Professional Documents
Culture Documents
Universiti Teknologi Malaysia User, Universiti Teknologi Malaysia, 14/08/2012 02:51, Uncontrolled Copy, (c) The British Standards Institution 2012
All rights reserved. Except as permitted under the Copyright, Designs and Patents
Act 1988, no part of this publication may be reproduced, stored in a retrieval
system or transmitted in any form or by any means – electronic, photocopying,
recording or otherwise – without prior permission in writing from the publisher.
Whilst every care has been taken in developing and compiling this publication, BSI
accepts no liability for any loss or damage caused, arising directly or indirectly in
connection with reliance on its contents except to the extent that such liability
may not be excluded in law.
While every effort has been made to trace all copyright holders, anyone claiming
copyright should get in touch with the BSI at the above address.
BSI has no responsibility for the persistence or accuracy of URLs for external or
third-party internet websites referred to in this book, and does not guarantee that
any content on such websites is, or will remain, accurate or appropriate.
ISBN 978-0-580-70782-7
Licensed Copy: Mr. Universiti Teknologi Malaysia User, Universiti Teknologi Malaysia, 14/08/2012 02:51, Uncontrolled Copy, (c) The British Standards Institution 2012
Contents
Foreword vii
1 Introduction 1
Market research today 2
The quality movement 4
Standards 5
Delivering quality 8
Improving quality 10
Foreword
This book draws on the experience and professional expertise that has
grown up around this subject in the UK over the last 35 years. The
authors have been closely involved over many years in the drafting and
implementation of research quality standards and their expertise and
understanding is displayed throughout this book.
1 Introduction
1
Throughout the text we will refer to ‘market research’ as shorthand for market, social and
opinion research unless explicitly stated.
2
Jackson, P (1997) Quality in Market Research, Kogan Page, London
1 Introduction
3
Research clients, buyers and users of research are all terms for an individual, organization,
department or division, internal or external, which requests or commissions research
projects.
4
Facebook’s factsheet, Facebook newsroom,
http://newsroom.fb.com/content/default.aspx?NewsAreaId=22 [accessed 14 May 2012].
These developments are innovative and exciting, but they can strain the
formal disciplines and in the forthcoming chapters we will consider the
impact of these trends on research standards and the various quality
schemes.
5
Fascinating Social Media Facts of Year 2010 – social media today,
http://socialmediatoday.com/soravjain/237864/fascinating-social-media-facts-year-2010
[accessed 12 April 2012].
1 Introduction
Quality has two broad and different senses. First there is quality as
excellence; the best; the superb. This usage is probably more common in
normal language, but as a working concept it has been either ignored or
rejected by management thinkers and the quality movement. Quality as
excellence, it is said, is just not a practical business concept. The second,
and the more applied, interpretation is relative quality; relative, that is,
to what is required. In other words good enough for the purpose in
mind: ‘fit for purpose’ and better rather than the best.
Standards
Standards
There are different types of standards: product or service standards,
quality process standards, professional ethical standards, business ethical
standards, etc. Whereas a product or service standard establishes the
criteria against which things can be measured in a verifiable and
consistent manner and process standards involve how they are produced,
others such as professional ethical standards are a form of applied ethics,
which concerns the moral issues that arise in a profession. Ethical
standards tend to be assessed by others within the profession (i.e.
self-regulation) as it requires the specialist knowledge of the profession
to effectively judge others’ professional moral conduct and behaviour.
1 Introduction
In the UK, market research was one of the first service standards to be
introduced (as BS 7911:1998, Specification for Organizations Conducting
Market Research) and codified the standards that the research sector had
already developed under its own steam, firstly via the introduction of the
Interviewer Quality Control Scheme (IQCS) and later by the activities of
the Market Research Quality Standards Association (MRQSA). Similar
initiatives were undertaken elsewhere e.g. Australia, Germany, France
and Spain. Chapter 8 will look at some of these early schemes,
particularly those standards for interviewers.
Standards
1 Introduction
Market research ethical standards are detailed within the various codes
of conduct, which set out the principles, rules and practices for individual
research practitioners. All the notable research codes are based on
comparable ethical principles. The International Chamber of Commerce
(ICC) and European Society for Opinion and Marketing Research
(ESOMAR) Code on Market and Social Research is the most widely used,
although most developed markets have their own Codes to reflect the
specific needs of their domestic market. The Code requirements are
discussed in Chapter 4.
Delivering quality
Given that quality has become a central concern to businesses and other
organizations, how is it to be achieved and delivered to the customer?
Three key elements are widely recognized although the emphasis varies
according to the writer, viewpoint and, arguably most important of all,
the nature of the organization seeking to deliver quality. These elements
are purpose, process and people.
Delivering quality
1 Introduction
Improving quality
The quality concept up to this point has been presented as static;
organizations seek it but then just need to keep it to a sufficient degree.
Clearly, however, this is not enough. Organizations need to seek quality
improvement on a continuous basis with the goal of ever moving
forward. The need for improvement is also, arguably, built into both
concepts of quality. Quality as excellence is always unattainable and the
best we can do is move a bit nearer the impossible. Relative quality is
also always dynamic since the requirements to be fulfilled are themselves
fluid and changing; in fact quality improvement is itself part of the
dynamic. Improvements raise expectations.
The nature and structure of the market research industry has important
consequences for quality. In this chapter we give a broad picture of the
industry including its boundaries and size, both overall and by segment.
The structure of the supply-side of the industry and of its workers is
described together with some specific consequences for quality issues.
How research is bought and used and trends in the industry are also
linked to aspects of quality and quality initiatives.
6
This figure is reproduced with permission by ESOMAR from their report.
14
Market research as an industry
market research
Figure 1 – Breakdown of global spend by research methods used in
Some would suggest that the size of the research market is significantly
undervalued due to limitations in the approach used for collecting
industry statistics. Many countries’ figures are based on associations’
estimates of their members’ activities as opposed to the market as a
whole, and there is currently no standard approach to determine which
organizations and activities should be included within the statistics for
research. For example, should data analytics activities (i.e. the process of
reviewing data to identify underlying patterns and relationships such as
advanced statistical modelling) or research into market information be
included? If these activities are conducted by organizations traditionally
associated with ‘research’ they tend to be included, but if conducted by
specialists or those not traditionally aligned with the sector they tend to
be excluded. Some countries have undertaken reappraisals of the size of
the domestic market to address some of these concerns and in the case
of the UK this resulted in a significant increase in the estimation of the
industry’s size. With the changing nature of research (to be discussed in
later chapters) there will be even more challenges to accurately measure
the sector size. The vagueness in these estimates is ironic given that
market researchers confidently make market size estimates all the time
for their clients. Despite its undoubted successes however, market
research, whatever its true size, is only a relatively small business sector
or industry, particularly when compared to other forms of marketing
communication such as advertising and direct marketing.
7
ESOMAR’s figures include activities conducted by market research companies or institutes,
and excludes work conducted by non-commercial institutes or research performed by
non-research companies that conduct their own research. Work sub-contracted to research
organizations outside the country is excluded in the domestic turnover figures to avoid
double counting.
8
The Market Research Society’s annual survey accounts for circa 40 per cent of the total
industry sales with contributors representing a diversity of company types and sizes, from
£100 million plus organizations to small sole traders.
The companies and organizations buying research are diverse not only in
terms of their own activities but also in size. However, as in most business
markets, a high proportion of the total research spend is by large
organizations with considerable budgets allocated to marketing activities.
Research suppliers
Some primary research work is carried out ‘in-house’ and in some areas
this is growing – particularly with the advent of easily accessible,
inexpensive research tools such as online survey and software packages
(the impact of which is to be discussed later) – but the largest proportion
of research is still carried out by commercial market research companies
or agencies11. Traditionally research companies were represented by trade
associations and the profession by professional bodies. Invariably one
body tends to be more dominant, sometimes due to historical issues, but
often it will be dictated by who is the regulator. The code-holding
association will tend be the dominant body. More recently there has also
been a movement to create sector bodies which represent both the trade
and profession and in some instances also the client. For example, in
Canada a totally new association was formed, the Market Research and
Intelligence Association, which included the membership of the former
professional, trade and client associations. In other countries such as the
UK, the larger body, in this instance the professional body – MRS – took
over the services formerly provided by the trade association, and changed
its focus to serve both the research business and the research profession.
The research market sector follows the classic Pareto principle.12 Of the
2010 global research revenues, the top 10 research organizations
represented just over 52 per cent of the total research revenues. This sort
of pattern is replicated nationally. Research companies are commercial
and profit-making businesses with varied ownership patterns. Of the
9
Based on figures from the ESOMAR Global Market Research Report 2011.
10
Based on figures from the ESOMAR Global Market Research Report 2011.
11
The research suppliers are a mix of companies, partnerships and independent consultants
but ‘research company’ will be used throughout for short-hand.
12
Sometimes known as the ‘80:20 rule’ that roughly 80 per cent of the effects come from 20
per cent of the causes.
Research suppliers
A wide range of market research services are available, but some research
companies will specialize in some way or other. One division is between
those offering ‘full service’ and those restricting operations to just part of
the research process (data collection and data processing being the most
common). Full service work includes the quality critical process of
research design – identifying client requirements, defining the research
problem and developing a suitable research programme using a range of
techniques and approaches to meet these objectives. On the other hand,
services such as ‘field and tab’ (these are organizations that just offer
data collection services i.e. go out in the ‘field’, collect data and tabulate
– ‘tab’ – the results) are limited to the more routine parts of the process.
The two types of company often inter-trade with ‘full service’ companies
(some of which have no internal data collection or analysis resources)
providing much of the turnover of ‘field and tab’ operations. Equally
some of the larger suppliers which have retained field and tab divisions
will offer data collection services to their competitors.
sector has clear benefits and quality implications. Certainly, when selling
to clients, research companies will always emphasize such knowledge and
buyers often put a premium on it.
13
Depth interview is an individual one-to-one research interview with a single respondent.
Depth interviews can also be done in pairs (i.e. a paired depth interview).
Market researchers
Market researchers
Market research is often referred to as a people business and this is true
in quite a few ways. Good research has depended on flair and
intellectual ability rather than technology, although this may be changing
to some extent with ‘intelligent’ IT tools. Placing business with a research
company is often about confidence in the individual selling the project,
in personal chemistry. Also – at least outside the largest agencies or
specialist data houses – the level of capital investment required in
research has been relatively low and a far more crucial issue for agency
management is recruiting and keeping good staff. However, while such
talented executive staff is vital, most people actually working in market
research are interviewers who are non-professional, relatively untrained
and are usually not employees of the research organization (most are
either workers or freelancers retained for specific projects). This particular
structural feature gives rise to some critical quality issues.
Entry-level positions
We shall leave a discussion of whether, or the extent to, which market
research is a profession until a later chapter but it is reasonable for now
at least to refer to professional market researchers. Typically recruits at
this level join a research company or in-house research department with
a first degree from university. Rather fewer have postgraduate degrees.
The disciplines from which they are drawn are very diverse; indeed this is
often the preference at recruitment (there is certainly a good sprinkling
from the social sciences, but also many with arts or natural science
degrees). Professional recruits also come from business courses but
generally the would-be researcher is not expected to have specifically
relevant research knowledge. This is usually acquired in the first two or
Professional-level positions
The archetypical job for professional-level staff in research companies is
‘research executive’, although different terms may be used depending on
the company and the level of seniority achieved. However, regardless of
title, the work undertaken includes many of the ‘thinking’ parts of the
research process:
The research executive also has a project management role to see that
the whole programme of work is carried out to specification and to
schedule. This side of the work will involve liaison with the ‘service’
departments of the agency – data collection and data processing – and
briefing staff within those areas on the specifics of the research. Finally
and at least as important as the other roles, the research executive also
14
This is based on in-house research with lapsed and former members undertaken by the
membership department at The Market Research Society based in the UK.
Market researchers
Client-side researchers
Professional-level staff working in research companies are within
communities that inevitably attach a high level of importance to
professional research values and implicitly to quality issues peculiar to
market research. People with similar training, background and
professional commitment to research also work for companies buying
and using the research output of agencies, i.e. client-side researchers,
who are often transitioning their careers between research companies
and clients. However, not all those buying research are necessarily
researchers. Many client-side research departments have been
rationalized or merged into broader marketing communications
functions, resulting in marketers having responsibility for market research
without necessarily having any real expertise or specialist training. In
addition, individuals in broader business functions will commission
research without any research knowledge at all. In conjunction with this
is the involvement of procurement as a specialist function, which has
meant that an increasing amount of research is being purchased by
procurement professionals. Many of these individuals have broad
responsibilities across a number of services and products, and therefore
Non-professional researcher
Professional market researchers are in the minority of those employed in
research; the majority are various other ‘non-professional’ staff. Market
research companies of course employ non-researchers for non-research
functions (e.g. finance staff), but the groups we really have in mind are
those involved in the primary data preparation and collection process,
particularly interviewers. The latter in turn are split between ‘field’
interviewers and phone unit staff. In the UK, circa 6,000 researchers are
members of the professional body but of these only approximately 200
are interviewers. This is in sharp contrast to the estimated total of 25,000
interviewers, of which around 5,000 carry out interviewing on a full-time
basis. The status and position of interviewers raise some particular quality
issues.
In Chapter 8 the quality issues arising from field and phone interviewing
are discussed in some detail. For the moment it is enough to mention
that field interviewers are recruited without any formal qualifications
(except basic qualifications such as literacy), receive only limited initial
training, work remotely from the designers of the research, cannot be
directly supervised for more than a very small part of the time and, as
essentially casual workers, commonly work for a number of research
companies (sometimes even in the same week). They also have limited, if
any, promotion prospects and the rates of pay offered are certainly not
high by most standards. Given this situation, company loyalty is unlikely
to be high. Even the local supervision of field interviewers may be by
staff who themselves often are working on a casual and freelance (if
better paid) basis for more than one research company.
15
The revenue growth rates in market research were until the 2008-09 global recession the
envy of many other business sectors.
It has been argued that the growth of the largest agencies has been or
will be at the expense of the medium-sized company – the organizational
type which is responsible for many of the established structural forms of
market research. However, medium-sized agencies have not been
squeezed out. What have emerged are medium-sized groups, created
through the acquisition of small and medium enterprises; these groups
are smaller than the large research groups but are of sufficient size to
compete, although they tend to specialize in specific sectors and/or
countries. At the other end of the scale new, small agencies continue to
be formed, particularly as a result of the 2008/9 recession and downturn.
As in the past, most of these will remain small ‘boutiques’ but others will
move up the scale.
16
Based on figures from the ESOMAR Global Market Research Report 2011.
17
Based on figures from the ESOMAR Global Market Research Report 2011.
Few would say that an economy Multi Purpose Vehicle (MPV) has the
quality, in the above sense, of a Ferrari, but if you want to carry your
children about or pick up compost from the garden centre it actually
better meets your needs. It matches a practical specification; it is ‘fit for
purpose’. Quality, in this sense, is relative to the requirements we have.
The MPV is much cheaper; quality as excellence has a cost which may be
too high for the intended purpose. At least in business decisions, we
cannot be carried away by intuitions of excellence and we need to take
account of what we spend on inputs with an effect on the end profit.
18
To pursue this aspect of quality further, Zen and the Art of Motorcycle Maintenance (Pirsig,
1999) is recommended.
So quality in market research is tied up with the output being useful. But
is being useful enough? We need to distinguish between outputs which
are thought to be useful and outputs which are actually useful. Good
decision making may be based on information but the decision still has
to be made; we cannot prevaricate forever. Market research can provide
a crutch so that some decision is made and is, therefore, thought to be
useful. It is a catalyst to a decision being made. But to be genuinely
useful the information produced through research also needs to be true
in some sense. In designing a research project we need to ensure that
data sought is relevant to specific marketing decision-making
requirements – to be potentially useful – but we also need to ensure the
methods are capable of getting at the truth. Quality in market research
is, therefore, bound up with truth and we next discuss the basis of this
truth.
some extent adapted the broad purposes and logic of the latter. It is,
therefore, worth thinking about how science seeks the truth and initially
we focus on the natural sciences.
One important point from Figure 2 is that the logic of the scientific
method is not based on induction – generalizing from a number of
observations to form a theory. Hypotheses to be tested may well be
developed in this way – there is no ‘correct’ way of thinking them up –
but testing proceeds by deduction from the theory to a specific
prediction that can be tested through observation. A hypothesis that is
incapable of leading to a testable predication is not valid science.20 Other
general observations about the scientific method are that much scientific
effort has gone and still goes into description rather than explanatory
theory and that conceptualization is very much part of the scientific
19
Natural scientists probably rarely refer to ‘positivism’; more often it is a term of abuse by
some schools of social science.
20
Cutting-edge physics may be the exception with nine dimensions and parallel universes but
is criticized by some on just these grounds.
work to test its validity for themselves as well as review the logic of the
chain from hypothesis, through observations made (the experimental
design) to the conclusions drawn; this process of testing is why
transparency is needed. Replication implies that the same results should
follow from the same methods regardless of who does the work; in this
respect at least the experimenter does not affect the results and science
in this sense is value free21.
Positivism in social science is not and never has been without its critics.
Whether or not positivism works in natural sciences (clearly whatever the
method used, hard science does work, as we all know in our daily life) it
is inappropriate to all or many of the tasks undertaken in social science.
The criticisms are many and varied. For example, it is claimed that the
objects of social science are just not comparable to phenomena in the
natural world and that value judgements (a no-no in natural science),
rather than being outside true knowledge, are largely the subject matter
of social science. Who does the work in much of social science is
inevitably relevant and this limits replicability. It is argued that the social
world can only be known through inevitably ‘slippery’ concepts, e.g.
‘market leader’, ‘brand share’, ‘consumer attitudes’ – and how these are
chosen and elaborated determines all else. The findings are limited to
what can be fitted within the conceptual framework chosen (though this
can be argued for the natural sciences as well). Additionally, social science
concepts have ‘meanings’; meanings to the observer who selects them
but also very often meanings to the observed, and both draw these
meanings from the social life they share. Even the language used to
define the concepts is based on social life. Value-free social science is a
21
Perhaps not wholly true in quantum mechanics – i.e. Schrödinger’s cat.
Most market research carried out does not even approximate to any
hypothetico-deductive methodology and induction largely rules.
Data-dredging approaches are, for example, how many researchers pass
much of their time and the output is prized by users. However, there are
other criticisms of the positivist position to consider. One is that most
market research, including quantitative, arguably does not even seek
‘explanation’. Description is considered quite sufficient. Projects
concerned with market structure analysis, user and attitude studies, and
customer satisfaction surveys, for example, are largely limited to
describing how things are rather than attempting to explain why they
are so. Even predictive modelling is mainly ‘empirical’ rather than truly
explanatory.
The study of language tells 20th century humans who they are ... the
disciplines dealing with language and other aspects of
communication hold the key to understanding human reality ...
marketing research must move towards an ever greater participation
in this linguistic turn.
O’Shaughnessy and Holbrook, 1988
All this is a bit rarefied and it would be wrong to suggest that the
average qualitative researcher spends his or her time worrying about
such matters. In fact most will have never heard of them at all. However,
that is not to say such theoretical concerns have no impact. They do, if
only through a slow trickle down. Researchers may well use, for example,
a semiotic framework without even knowing the term. Also, most users
of qualitative research will judge the outcomes in only the fuzziest way
and rely on their confidence in the practitioner. Good qualitative research
is widely regarded to be that done by good qualitative researchers; it is
inevitably subjective and you either accept this and its consequences or
you reject such research. One consequence is that two first-rate
qualitative researchers working to the same brief may well come up with
two quite different sets of findings (this is also an argument for having at
least two researchers on a project). However, quantitative research also
has an arguably subjective element (including in questionnaire design)
and, therefore, it is not so simple a matter as good and valid quantitative
research versus unreliable and dubious qualitative work.
Not so long ago, qualitative research was almost synonymous with focus
groups (group discussions) supplemented with occasional depth
interviews with small samples (rather more costly than groups). More
recently, a wider range of techniques are in vogue including groups and
interviews online rather than face to face, ethnographic studies (really a
revived old methodology, originating in social anthropology), and
monitoring and ‘data scraping’ online social media (e.g. Facebook). None
of these, however, change the more fundamental questions of qualitative
research and how the truth is found.
Addressing validity
Validity relates to the subjects of any research study and the fundamental
question here is whether what we think we know about them truly
corresponds to how they really are, act or think. It would be a simple
matter to establish validity if we could in some way lay down the
observations made or measurements taken of our subjects alongside
them but we cannot do this. We can only know the subjects in terms of
such observations and measurements and in an absolute sense, what is
true, over and above any observations, is unknowable. The best we can
do is to increase validity through using effective tools (effective on the
basis of both theory and practice) in the belief that good work comes
from the successful application of these tools.
Definition of subjects
While much has been written on the measurement and observation tools
of research, there is surprisingly less apparent concern about the need to
use these among relevant research subjects. It is no use having the
perfect questionnaire if it is asked of respondents who have no
involvement in the decisions being studied or whose attitudes or beliefs
are quite irrelevant to the objectives of the research. Ask the right
questions certainly, but ask them of the right people. The right
respondents may be clear-cut but in some projects insufficient thought is
given to the problem; pilot research may be needed to provide a valid
solution before the main stage of fieldwork begins and this is seldom
undertaken. Business-to-business research is an area where respondent
qualification is often a special concern. In qualitative research the
question of whether the respondents are ‘right’ is usually transparent in
groups or depth interviews but in larger scale quantitative surveys the
extent of any problem may remain hidden.
Addressing reliability
While validity is ensured at the level of the research subjects, reliability
addresses the other major problem in arriving at the truth: on what basis
can we generalize the results to a wider population, which raises issues
of inference. In quantitative research the key to this is sampling and rules
of statistical inference, but there is no such obvious solution for
qualitative researchers. Few clients really want to know the opinion of 20
respondents, no matter how skilfully their deepest feelings have been
probed, unless they can be taken as representative of some wider group
such as all consumers of a product group, or those from a particular
social class.22 Even if the researcher does not formally extrapolate to a
wider group, this is likely to be done by others including the client.
Alternative bases for inference in qualitative research usually rely on
some loose assumptions that, even though the sample is small, it can be
taken as representative, providing the measures or observations extended
are not claimed to have any precision. These problems have always been
the bugbear of qualitative research and, to some, imply that such
methods should never or hardly ever stand alone.
22
In the UK, research has its own social class grading system which is primarily based on
occupational grouping; often in media articles you will find references to ‘ABC1’s’ to
describe individuals that are from the upper end of the social demographic scale with C2,
D and E at the other end of the spectrum.
Sampling
We will not enter into any exposition of either the theory or problems of
sampling within this chapter or even book. Sampling is the basis of
inference in quantitative research and the validity of any conclusions
drawn from the sample data to the general population (from which the
sample has been drawn) depends on the rigour of the sampling
techniques used. Sampling is based on statistical probability theory and
the necessary range of sampling error in any research has to be both
recognized and taken into account in drawing research conclusions.
Sampling error itself does not, however, compromise the pursuit of truth
and quality. It does, however, qualify it. The effective sample and sample
size is of course the achieved sample. Non-response is always a potential
problem and, with falling respondent cooperation levels, of increasing
concern (particularly as those that do respond generally are those with
more time e.g. the unemployed, young mums, retired, etc.). The
self-selection bias of self-completion research – whether by mail or online
– has always been seen as a limitation of that sort of research. However,
with phone interview response rates down to around 10 per cent of
those initially contacted, self-selection bias also affects administered
interviews.
Repeatability
External consistency
How market research fits into the wider methods of social science and
the scientific method more generally is a debate which will not end, at
least not in the foreseeable future. There will continue to be divergent
views. Much of market research is not ‘scientific’23 in the more rigorous
use of the term. Despite this, professional researchers need to think
about the issues and have a vision of how truth is reached. Good
research depends on good methodology and whatever research users
may say, the need to develop effective methods and critically debate
them has to be a continuing professional concern. Market research
projects need to be fit for purpose and, although this has to be linked to
costs and whilst high accuracy levels may not always be required, below
some level methods may be simply inadequate to produce usefully true
results. However, as in natural and social sciences, there is no final and
fixed list of valid methods. Many become well established and some
principles cannot be breached lightly but methods and techniques evolve
as part of a progressive discipline. In selecting methods the researchers
need to know what has been shown to work where and what has been
proven invalid – this is part of professional education and development.
23
In Germany market research is required to be carried out to ‘scientific principles’ but this
has a wider meaning than is common in English usage.
One vital aspect of science that is directly applicable is the need for
transparency and replicability. Except rarely, the output of commercial
research tends not to be published in journals. However, the need for
users of the research to understand how the findings have been
produced and to be able to later replicate as well as evaluate them is still
vital. Even where a research user or client disdains to understand the
mechanics, professional research practice requires methodological details
to be documented and at least available on request.24
24
Proprietary research techniques and commercial advantage may sometimes raise issues
about limiting transparency.
One of the problems with this model for the market research sector is
generally that the professional bodies and not the trade associations are
the code-holders and regulators. The result is that the professional bodies
can exercise no direct control over research suppliers, and there is
inevitably overlap in both interests and activities and some potential for
conflict.
25
Financial & Operating Ratios for Professional Bodies: 2009/2010, Professional Association
Research Network (PARN) reported that in 2010 57% of professional bodies increased their
membership base whilst 24% saw a decrease and 19% no change.
The regulatory issue is not currently pressing in the UK but this could
change without any apparent cause. Online research, for example, could
trigger new concerns arising from more general issues about privacy in
the online age.
Professional ethics
Being a professional is supposed to involve more than merely following
an occupation or earning an income; a profession is supposed to be a
vocation or a calling. This implies that individuals would want to pursue
the activity regardless of reward. This is clearly not the case today; in
general, people become professionals to enhance their income, not
restrict it. Market research is also unlikely to be carried out purely out of
a love of the subject or for humanity’s sake. However, the concept of a
vocation or a calling still has some force and implies that even though
members of the profession may earn lucrative salaries, reward is in the
context of commitment to norms and standards set by the profession. A
professional is expected not to be motivated solely by money, even if it is
allowed within the framework of the general law. These restrictions are
supposed to protect clients and provide some guarantee of quality;
professionals are expected to deliver the best and the truth, or at least
strive to do so. This sense of calling is implicitly recognized within some
sectors of market research, particularly those where ‘craft’ still forms its
core. However, many professions, and research is no exception, are
experiencing a movement away from craft towards industrialization and
mass market production, from bespoke craftsmen making bespoke
products to a ‘flat pack’ approach by non-specialist, cheaper and less
experienced labour. In research terms this can be the difference between
some ‘cheap and dirty’ online data collection, and the large scale
longitudinal social research project, with the vast majority somewhere in
between this spectrum.
Professional organizations
researchers, they are not always open to what is required to achieve this,
nor for the control to be relinquished to enable it to be gathered.
Professional organizations
Professions are invariably organized as formal bodies with constitutions,
codes, management, and services to members and the public, and they
may own substantial assets. In the UK some bodies have submitted
themselves to greater scrutiny to obtain Royal Charters e.g. The
Chartered Institute of Marketing and The Chartered Institute of
Professional Development. However, organization and structure alone do
not define professions and some bodies cannot claim such a status. Some
professions have a number of competing bodies but in the UK MRS is the
only substantial body covering the market research profession. That is not
to say there are no other research bodies in the UK – there are numerous
bodies covering client-side researchers, social research, qualitative
research, business to business research, independent consultants, account
planners, etc – but MRS is the body which covers the broad church of all
those involved in research from whatever background or interest.
MRS is also a good example of a professional body that has developed its
business model to be less reliant on membership (currently membership
subscriptions represent less than 20 per cent of its total turnover) by
cultivating its commercial activities e.g. trade publications such as
Research magazine and the trade directory, Research Buyer’s Guide, and
their associated websites, www.research-live.com and
www.researchbuyersguide.com. This has the advantage of diversifying the
business base with the disadvantage that the commercial activities tend
to be discretionary (sponsorship, advertising, training and conference
53
Licensed Copy: Mr. Universiti Teknologi Malaysia User, Universiti Teknologi Malaysia, 14/08/2012 02:51, Uncontrolled Copy, (c) The British Standards Institution 2012
Professional bodies usually have formal codes of conduct which set the
ethical principles and rules for individual practitioners; and trade
associations have codes of practice which set corporate rules. Such codes
may be referred to as professional standards but there is a difference
between codes of practice and standards. Codes are also different to the
constitutions of the professional bodies. See Figure 5 for the differences
between these three types of document.
Key principles
The current MRS Code of Conduct (2010) contains 10 overarching
principles which are applied and interpreted via 74 rules split between
the responsibilities of the researcher and the research process. A
one-page summary of the MRS Code of Conduct key principles is
reproduced below.
Key principles
ethical rules of the MRS Code of Conduct, and provide a framework for
all professional activities that a researcher may be engaged in as opposed
to just exercises for research purposes.
The key principles of the MRS Code of Conduct reflect the broadening of
researchers’ activities and apply to when researchers are conducting
research and non-research activities. In relation to clients, the MRS Code
generally avoids impinging on commercial considerations, although it
does stipulate the need for contracts between all parties involved in a
research project (research supplier, subcontractors, clients, etc.) as
misunderstandings in this area often give rise to complaints. Mirroring
respondent confidentiality, however, are requirements for keeping
information about or supplied by the client confidential and this includes
keeping even the identity of the client confidential (e.g. from
respondents); the fact that a particular company has an interest at all in a
market or product may be commercially sensitive. For various practical
reasons, quite apart from the client’s interest, researchers may prefer not
to identify clients. However, respondent data protection rights may
require client disclosure in some circumstances e.g. when respondents are
contacted from the client’s own database.
same with all other major research code holders, including international
bodies such as ESOMAR. Insofar as research codes and associated
disciplinary procedures impinge on quality of research, it is not a very
effective vehicle for giving clients assurance about quality levels and, to
be fair, this is not a primary purpose of ethical codes.
of design, using this term in a broad sense (e.g. a solicitor planning the
conduct of a case is carrying out ‘design’ work).
Is professionalism enough?
Market research by most criteria is a profession, even if the wider world
does not always recognize it to be one. It has a well-established
professional body and high-level of intellectual content. Within the
profession there is some consensus about good practice and therefore
about what makes for quality, or at least quality as excellence. This
consensus is also enough to provide a body of knowledge which it is
agreed should be acquired by new entrants. However, there is far from
unanimity on what constitutes good practice, not only in relation to
specifics and details but even the basis by which the ‘good’ can be
determined. Furthermore, the ability of the profession to ensure that
quality, whatever its makeup, is delivered in practice is limited for
structural reasons. Market research is very much an open profession and
not all practitioners are within the professional body (nor is it reasonable
to assume that those outside are necessarily inferior in skills). Even full
membership in a recognized professional research body does not
guarantee any particular level of competence or ability to deliver quality
(although in future this might change). There is also the problem that
the level of control exercised by codes of conduct and enforcement is
quite limited. Similarly, research services may be delivered by research
suppliers who are not members of code-holding bodies (this will vary
between countries) and supplier organizations may not be managed and
overseen by those from within the research profession. Increasing
Is professionalism enough?
Satisfying clients
That market research companies – or any business – must over a period
of time satisfy clients or customers is almost a truism. No client is bound
forever to any research company and if seriously dissatisfied the client
will sooner or later take their business elsewhere. Clients expect value for
money, research which meets their purposes and which can have a
demonstrable and measurable return on investment. If the research
company concerned consistently performs poorly and fails to meet clients’
internal requirements it will soon be out of business. Like any other
business, therefore, relative quality – satisfying clients – has to be a major
objective whether or not it is set out in, for example, a ‘quality policy’.
client will recognize the underlying need for ‘true’ research. On the other
hand, a pursuit of excellence, irrespective of client wants, has its own
problems. For one thing, if the satisfaction of manifest wants is ignored
there is a good chance that business will not be placed; the proposal will
simply be lost if it inadequately addresses stated wants, even though the
research proposed may objectively be the perfect solution. Also – and
many research companies have taken a long time to realize this – clients’
wants go beyond areas that correspond to any objective research needs.
Confidence in the people involved in the research work and whoever sells
it, clarity in all forms of communication relating to the work, and high
quality reporting are all likely to be client wants that have to be satisfied.
Commercially, it is such considerations that often decide who will have
the business. Finally, wants must be balanced with needs within a budget;
the cost of ‘best’ research may be unrealistically high in relation to the
client’s available budget or in relation to the value of the information
provided.26 It may not even be possible to meet the client’s needs at all
within the budget proposed; the affordable research methodology may
just be grossly inadequate. In most cases, however, some compromise or
trade-off can resolve the dilemma.
Satisfying clients
Conformity to specification
Once identified, client requirements at the individual project level can be
set out as part of a specification. This will also include some statement of
the outputs which will be delivered to meet these requirements. The
most common tangible form of such a specification is a research proposal
(at least it is in ad hoc research). However, the ‘full’ specification of what
is required and will be delivered is always greater than what is normally
covered in a proposal document. For one thing there may be an implicit
assumption of professional standards of practice beyond what is formally
set down, and this may include that any methods used will conform to
generally recognized and valid practice. There are also the requirements
that relate to service outside the area of objective research practice –
good communication, clarity in reporting, timeliness, levels of client
contact, and involvement. Some of these may be covered explicitly in a
proposal while others may be set out in promotional literature or
standard operating procedures or as part of a formal quality policy.
Agreed and recognized industry or quality standards may also apply.
However set out, and even if part of it is implicit, a specification for the
research project will exist and should be recognized by the supplier and,
as far as practical, explicitly agreed with the client. Difficulties arise and
serious quality problems result where this specification leaves out some
important requirements, since there is no possibility that the research
company can satisfy the client if they are unaware of what is required. If,
however, a more or less complete specification is achieved, quality
becomes a matter of meeting that specification: meet it in all important
particulars and quality is achieved; fail to do so in one or more areas and
the service provided is less than high quality. Here, then, is another sense
of quality – conformity to specification.
Bringing money into quality is usually by some caveat of ‘least cost’. For
example, customer requirements will be met at least cost, or conformity
to specification will be at least cost. Least cost is regarded as the mirror
of most profit although this congruence is not, strictly speaking, logically
watertight. In one sense, least cost can compromise what is delivered to
the customer; a budget airline for a long haul flight may be the least cost
option but for a comfortable flight, convenience and all the added
extras, business class might be more appropriate. Similarly an online
sample of 500 may be cheaper than a random face-to-face sample of
1,000 but may be inadequate to meet the research objective and
purpose. However, least cost is a valid concept if considered
post-specification. Once what will be delivered is defined and agreed,
least cost becomes a matter of efficiency.
Achieving efficiency
There are three main areas to seek efficiency in market research: in
methods of working, in productivity, and in minimization of errors.
Working methods
27
Adam Smith in An Inquiry into the Nature and Causes of the Wealth of Nations (1776),
described the production of a pin divided into eighteen distinct operations and used this
to describe how efficiency could be increased through the division of labour.
It may appear from the above that we are suggesting that various forms
of non-specialist flexibility are always better than a pronounced division
of labour in market research, but this is not our intention. Division of
labour is always a means to be considered in the pursuit of efficiency
providing that any problems it creates do not outweigh the realizable
benefits. Some research companies may paradoxically have too much and
too little division of labour, with tasks that need to be integrated divided
and tasks that would be better divided, integrated. The changing nature
of research, particularly how data are gathered – less face to face and
more indirect via digital techniques28 – means that the ‘old’ divisions
between field, operations, head office, etc. may result in entirely
different divisions of labour.
28
Based on figures from the ESOMAR Global Market Research Report 2011.
Productivity
Minimization of errors
measured. All sorts of measures are possible within the processes (e.g.
error rates, instances of non-conformity to specification, production costs
and productivity) or of the outputs (e.g. customer complaint levels,
customer satisfaction, market performance, financial performance); some
of these measures are more direct than others, which may reflect
‘outside’ factors independent of quality. Market research techniques
themselves can be a quality measurement tool, particularly with customer
satisfaction research.
responsibility also works very well at the people level through the strong
emphasis on professional development, both initial and continuing, and
the advancement of professional knowledge (professional bodies in their
role as learned societies).
Quality control also has some fairly obvious limitations even in relation to
process. For one thing, scrap represents waste and unnecessary costs but
within the model nothing is done to address why scrap arises or why the
rates go up or down. The output of quality control of this sort is
inefficiency and with the underlying problems not considered, no
remedies can be sought or found. Also, who is to say the ‘policeman’ is
up to the job or that the criteria by which products are judged
correspond to requirements. In market research the application of the
quality control model in this form is clearly of little value since the
quality of research can scarcely, or only very partially, be inspected in any
final output. A check of a written report may pick up processing errors,
poor layout or even an inadequate coverage of the research objectives,
but even if it passes those checks it still does not say the work is any
good; the research design may be poor, or badly executed, the data
collection haphazard or full of errors and the analysis may have made
everything worse.
As with many other sectors, quality assurance has made an impact on and
inroads into market research, with companies implementing the most
widely adopted standard for quality assurance – the BS EN ISO 9000 series
29
We use the two terms – quality assurance and quality management – as synonymous
although arguably quality management is the means of achieving quality assurance.
A point to note about both quality control and quality assurance is that
the specification of what should be attained in quality terms comes from
outside and particularly from customers and clients and their
requirements. Quality control takes this as given while quality assurance
should incorporate methods for establishing or checking what these are.
However, quality assurance has no built-in standards of product or service
quality. It is a method for ensuring that such standards, whatever these
are, are identified and met constantly or consistently, but it is not a
method of defining them. As an aid to both customers and suppliers, it
may be found useful to define requirements and the specification wholly
or partly in terms of some recognized and independent standard. Quality,
or at least minimum quality, then becomes a matter of conformity to the
standard. In manufacturing, quality standards, particularly BS and ISO
standards are very well established. Service standards have also become
increasingly prevalent in recent years and in the UK, market research led
the way in this respect with the first true BS service standard
(BS 7911:2003, Specification for organizations conducting market
research). In turn UK standards for market research influenced the
development of comparable published standards in other countries and
eventually led to the international research standard ISO 20252, first
developed in 2006 and subsequently updated in 2012, and its
accompanying standard for access panels: BS ISO 26362:2009, Access
panels in market, opinion and social research – Vocabulary and service
requirements (see Chapter 9).
Some would also contend that TQM sets up artificial limitations to quality
assurance so that the new methods can claim to surpass them. Also,
effective quality assurance programmes recognize that customers’ wants
and needs change and shift with time – partly because of the efforts of
suppliers to give new experiences – and that quality, therefore, is not and
cannot be static. However, TQM cannot be dismissed, if only because of
its level of uptake. Also it is not necessarily a competitor of quality
assurance; it can be an extension. The standard BS EN ISO 900430 is
intended to provide a bridge between the quality assurance approach of
BS EN ISO 9001 and a more TQM approach. Also some regard quality
assurance as necessary underpinning on which further progress can be
built. Others, however, are more sceptical; at least to quality assurance
built around standards such as BS EN ISO 9001, which some perceive as
not necessarily suitable for their business needs.
30
Although BS EN ISO 9004 is classed as a standard it is better considered as a guidance
document on quality management and it is explicitly stated by ISO that BS EN ISO 9004 ‘is
not intended for certification, regulatory or contractual use’; organizations are not
assessed externally against this standard.
Business re-engineering is clearly not at the cuddly end of TQM and the
reality may often involve acute pain for the participants, with
down-sizing (i.e. fewer staff undertaking more of the processes and
having to take greater responsibility for activities) and de-layering (i.e.
making management staff redundant). These aspects may clearly work
against commitment and be seen as anti-people, although they may well
improve shareholder value regardless of staff morale (which is not, after
all, an end of capitalist enterprise). Moreover, with developments in
technology – and more processes being driven by technology rather than
people – the cycle of re-engineering, re-structuring, resource rationing,
etc. is now a permanent feature of business life including market
research. Although the sector has experienced an unprecedented period
of growth (only recently stemmed by the global recession) there are
generally fewer traditional researchers employed within the sector who
are generating more research than ever before.
formal business quality standards such as TQM. In practice, TQM may well
take the form of a linking and amalgamation of professionalism, quality
assurance, and the adoption of industry minimum standards such as ISO
and IQCS together with softer, people programmes (several research
companies have taken up IiP). As an industry, little effort is needed to
persuade market researchers of the importance of people to achieve
quality. Techniques such as empowerment, facilitating management and
the learning organization are already often in place in market research
companies and were so before the terminology was invented.
31
It should be noted that the interpretation of legislation and its implications for research
detailed in this chapter is not, and should not be regarded, as a legal document.
32
Special Eurobarometer 359 Report – Attitudes on Data Protection and Electronic Identity in
the European Union (June 2011).
The legal basis for the protection of personal data originates in the
Universal Declaration of Human Rights of 1948 in which Article 12 states:
This was reinforced with the Council of Europe’s 1950 Convention for the
Protection of Human Rights and Fundamental Freedoms in which Article
8 provides for the right to respect for private and family life.
Europe has led the way in developing data protection legislation. In 1981
the Council of Europe adopted the Convention of the Protection of
Individuals with regard to Automatic Processing of Personal Data (known
as Convention 108) and this lead to the first comprehensive, national
data protection laws. Convention 108 was built on initial legislation
which had developed in countries such as France and Germany during the
70s. The national laws were originally divergent and to address this, the
European Union (EU) introduced the Data Protection Directive (95/46/EC),
with the aim of harmonizing European data protection legislation. As
membership of the EU has grown so have the number of countries that
have developed privacy legislation (a requirement of European
membership is to adopt Directives) and increasing numbers of
non-European countries have taken a similar path (Australia, Canada,
Japan, Philippines, South Africa, Israel, etc.), as have significant economic
regions such as California in the US.
The ‘rules’ for data protection are defined in Article 6 of the Directive.
The key points from Article 6 are:
The issue of consent is even less clear cut in some online e-ethnography
projects where researchers observe behaviour of individuals in social
media networks and forums. The ‘observer effect’ whereby subjects
modify their behaviour in response to the fact that they are being
studied, is an important research design consideration. Nevertheless the
researcher cannot ignore the requirements of the legislation, however
tempting it might be for researchers to remain unobserved; ‘lurking’ on
such interactions. Legally, if personal data are to be collected for research
purposes, data protection legislation must be adhered to, and researchers
would need to identify themselves and state the purpose of their
observation including what they intend to do with any data collected.
Although this is a legal requirement when collecting personal data, it
would still apply irrespective of whether personal data are collected due
to ethical code requirements. The main research codes all contain
principles based on the concept of informed consent, respecting
respondents and avoiding respondent harm. To lurk in an online social
forum, without correctly identifying yourself as a researcher, would be in
33
Ohm, P (2010), Broken Promises of Privacy: responding to the Surprising Failure of
Anonymisation.
The Directive also gives rights to data subjects, one of which is the right
to object to direct marketing. The Privacy & Electronic Communications
Regulations 2003 enables individuals to exercise this right via telephone
direct marketing. Do not call/preference lists operate in most European
countries where individuals can register their personal (and in some
countries business/corporate) telephone numbers, with the purpose of
opting out of being contacted for direct marketing purposes. Although
market research is exempt from having to screen against most of these
lists (e.g. the Telephone Preference Scheme in the UK, the Do not Call list
in the US, etc), as individuals are only opting out of direct marketing,
some research organizations and clients screen research samples against
them. This is partly in the belief that subscribers to such services are less
likely to agree to participate in research projects (although this is not
always the case) and to reduce the likelihood of respondent complaints
and dissatisfaction. However, such decisions introduce a bias (or another
bias) to the sampling methodology by excluding potential respondents;
subscribers to opt out lists, for example, tend to be biased towards the
higher end of the social demographic.
In other environments this right is enacted via the use of so-called ‘Fair
Processing Notices’ where a data subject can opt in or opt out of direct
marketing (the ubiquitous tick boxes found at the end of any paper or
online application forms). The emergence of this concept is noteworthy
for research in several respects. Some of these statements can include
opting in or out of research in addition to direct marketing, although
there is no legal obligation to do so. This is sometimes because the data
controller has elected to be more stringent than the legal requirements,
but often it is due to poorly worded notices, together with a weak
understanding of the legal requirements. Either way, if research has been
included in the wording of fair processing notices, the wishes of the data
subject must be respected, and any respondents which have elected not
to be contacted for research purposes must be screened out of research
samples.
The same can apply with passive data collection techniques, online
technology may enable researchers to gather large amounts of personal
data relatively easily (by using data scraping techniques which are
software programmes which enable vast amounts of personal data to be
‘scraped’ off internet such as social media sites), but personal data should
only be collected if it is to be actively used for research (and again this
should only be done with the consent of the individuals whose personal
data are gathered). If the personal data are not going to be used for a
legitimate research purpose, it should not be gathered as to do so would
be excessive and not relevant, and hence in breach of data protection
legislation. Ultimately research projects should always be designed with
the minimum amount of demands being placed on respondents and so
any legislative requirement which stifles boring, repetitive and
unnecessary research questioning should be welcomed.
34
The Data Protection Act 1998 & Market Research: Guidance for MRS Members (September
2003) first introduced this idea and this has been adopted elsewhere.
tell them. The key requirement is that researchers accurately record what
respondents tell them and do not inaccurately attribute information to
the wrong respondents. The requirement for keeping data accurate and
up to date also has practical implications in the management of research
panels, where processes need to be in place to ensure that panellists are
frequently contacted (to make sure they are still active) and to ensure
any recorded information about the panellists is accurate and frequently
updated. For simple ad hoc projects the requirement for accuracy applies
but there is not a practical need to keep the data up to date. As for data
retention this will usually be determined by the requirements of the
research e.g. a longitudinal study will usually require longer retention
periods, ad hoc much less although retention for some period for quality
(and accuracy) checks is needed. In the absence of specific contractual
requirements, BS ISO 20252 sets out practical periods for data retention
(one year for primary data, such as questionnaires, and two years for
secondary data such as research reports).35
Data security
All controllers must implement appropriate technical and organizational
measures to protect personal data against accidental or unlawful
destruction or loss, alteration or unauthorized disclosure or access. In
today’s technological environment this is one of the toughest challenges.
35
There are also some limited exemptions within the Directive for following requirements
such as data retention for data collected for statistical and historical purposes but this has
limited coverage to only some statistical quantitative research.
With so much research data flying around some is bound to get lost, and
the more handovers of data between parties there are, the bigger the
risks. A very simple research project would involve at least three of four
data handovers between client, researcher, sub-contractors and possible
interviewers (see Figure 8 for a ‘typical’ very basic research data flow).
The client’s security is likely to be good, so are most researchers, but
what about an interviewer? Someone who is not an employee of a
company, probably works from home and is not likely to have Fort
Knox-type security arrangements – how do you keep the data secure
there? Or outsourcing arrangements where research data processing and
collection is conducted in another country where direct and frequent
monitoring of arrangements might be difficult? This is without even
considering multiple sub-contractors, multi-country projects where the
risks can rise ten-fold. Reports of general data loss incidences are
increasing. The British government seem particularly prone to losing vast
quantities of their citizen’s data; two CDs containing 25 million personal
records of families of child benefits claimants went missing in the post in
2007, a hard disk with 3 million details of candidates for the driving
theory test was also lost in 2007 by US subcontractors working on behalf
of the Driving Standards Agency and multiple smaller losses have
subsequently been reported, some of which included research data being
lost (mercifully these are currently few and far between).
Freedom of information
Online e-privacy
Other pieces of legislation have been introduced with the purpose of
addressing data protection and privacy issues but in specific contexts. For
example the European Directive (2002/58/EC) ‘concerning the processing
of personal data and the protection of privacy in the electronic
communications sector’ (Directive on privacy and electronic
communications) – called the ‘e-Privacy Directive’ for short – aims to
address data collected by invisible processing. There are a number of
techniques used by researchers which employ invisible processing
techniques – the main ones are cookies and digital fingerprinting.
A cookie, of the non-food variety, is a piece of text that a web server can
store on a user’s hard drive with the purpose of storing information and
later retrieving it. Researchers use cookies for a number of reasons: to
ensure that respondents only participate once in any given research
project; to protect against unauthorized intrusion into protected pages
and/or sites; to protect against fraud, etc. Similarly, digital fingerprinting
is a form of technology which invisibly and automatically collects
individual’s computer browser and hardware settings for the purposes of
producing a unique ID for a computer. The benefit for research of such
technology is that it helps to identify potentially fraudulent respondents,
particularly panellists who try to join research panels using multiple
identities through the same computer.
These are clear research quality control reasons for taking advantage of
these technologies, but it must be done in adherence with the
requirements of the e-Privacy Directive. The main requirement of the
Directive is that such technology cannot be deployed without obtaining
the consent of the user, having provided them with clear and
comprehensive information about the purposes of the storage or access
to the information that is being collected. This could be achieved by
providing, within research preambles, some top line cookie information
and why they are needed, providing a link through to more detailed,
technical information about specific types of cookies used, how
information is stored, who would have access to the information, how to
disable them, etc. Similarly when using digital fingerprinting technology
in the management of a research panel for example, to explain within
the terms and conditions for the panel that such technology is used, for
what purpose, etc.
Freedom of information
It is also worth mentioning the other side of the data equation –
freedom of information. Whereas the focus of privacy legislation is to
protect personal data and the rights of the individual to have a private
life, freedom of information is all about promoting openness by enabling
Other communications
So when it comes to legislation should researchers only be worried about
data legislation? Well, no. There are various strands of legislation which
impact on the technology that research uses and/or the environment in
which research tends to be undertaken.
legislation can be found elsewhere in Europe and the US. The primary
focus of the UK legislation is to regulate the medium e.g. automated
diallers; whereas other countries such as the US have focused on both the
medium and the purpose e.g. diallers used for telemarketing, plus
restrictions on the use of auto-diallers when calling mobile phone
numbers. The difficulty with legislation in this area is that the traditional
argument that market research is not a commercial activity, and hence
should be given special treatment, is not necessarily sufficient (not in the
UK context for example). Many consumers who receive such calls perceive
them as a nuisance irrespective of whether they are generated by market
researchers or direct marketers. The legislation therefore places the same
restrictions on all users of such equipment irrespective of intended
purpose.
It is not just the telephone and online research environment where legal
restrictions can be encountered. On the ground, conducting research in
shopping areas or in-street can be just as problematic. Local laws
empowering communities to create ‘No Cold Calling Zones’, closed and
gated communities, or commercial shopping districts restricting any
non-retail related activities (such as street interviewing or fund raising)
can all make face-to-face research interviewing difficult. Often the focus
of these restrictions is not to stop research interviewing, nor are they
necessarily underpinned by sufficient legislation to enable residents or
town centre managers to stop interviewers from doing their job, but the
end result is that these restrictions curtail what researchers can practically
do.
36
Then and now, BS EN ISO 9001 is part of the BS EN ISO 9000 series of standards. Currently
the series includes, as well as BS EN ISO 9001, BS EN ISO 9000:2005 which provides
definitions, introduction to the concepts behind the standards and some guidance on
application of the standards. There is also BS EN ISO 9004 which provides guidance on a
more TQM approach (mentioned in Chapter 5). Withdrawn standards in the series include
ISO 9002 and 9003 which were similar to BS EN ISO 9001 but with more restricted
application.
BS ISO 20252 requires little extra implementation but not vice versa;
quite a lot extra may be required to extend BS EN ISO 9001 certification
to also cover BS ISO 20252.
BS EN ISO 9001 was a major input into the development of national and
later international standards for research. First of all, it brought to the
attention of the research industry the concept of national and
international standards and standard bodies. In turn this stimulated
debates about quality in market research; what it means and how it is
achieved. One view was that the quality management approach behind
BS EN ISO 9001 could very usefully be applied to all the key market
research processes (although many may not have been aware of it, the
quality management or assurance approach was already used in
fieldwork standards, such as IQCS (see Chapter 8) in the UK). At the other
end of the spectrum, some did not like what was perceived as a ‘box
ticking’ culture displacing real quality (i.e. quality as excellence).
Another view, and in the end the dominant one, was that an
BS EN ISO 9001 approach was basically sound but a research-specific
variant or new standard was needed as well. As will be discussed later in
this chapter, BS EN ISO 9001 underpins quality assurance but does not
itself require any specific quality level; a research company could ensure
quality consistency but work well below recognized good research
practice, e.g. interviewer validation levels, although this would obviously
jeopardise any business in the longer term. The upshot was a consensus
to develop new quality standards for the whole market research process
incorporating BS EN ISO 9001 concepts and coupled with developing a
similar style of assessment and certification regime. Because of both its
historical and ongoing relevance to the industry, a book on quality in
market research, therefore, needs a chapter on BS EN ISO 9001; an
understanding of what the standard is (and is not) and how it can be
implemented by a market research company.
37
The specimen procedures are to meet BS ISO 20252:2012 but the format is equally relevant
to BS EN ISO 9001.
The standard also refers to the need for records. Critics of BS EN ISO 9001
have claimed that the consequences of implementation is a bureaucratic
explosion with a proliferation of records; a form for everything and a file
for every form! There is some truth in this but only where quality
management systems have been badly planned. There is no requirement
for records just to meet the needs of the system itself38 or to keep
assessors happy. Depending on the processes there will be some need to
record what has been done as part of the process but this need should
exist regardless of implementing the standard. If, for example,
interviewers’ work is validated in systematic ways (e.g. by re-contacting a
proportion of respondents) some record is needed to show the agreed
number of interviews are validated, which interviewers’ work was
covered, what was found out and what was done about any problems
identified. BS EN ISO 9001 requires quality improvement to be an aim,
including in systems, processes and outputs, and records of where things
have gone wrong, are an essential starting point for making positive
changes. In market research work, documents of one form or another is
the physical output of the activity and often these themselves provide
the record of the process; a form is not needed to show a questionnaire
has been drafted, the questionnaire is its own record and its conformity
to such as a standard format is there to be seen.
Management responsibility
Companies implementing BS EN ISO 9001 or similar standards generally
have a quality manager responsible for the day-to-day administration of
the system. The standard uses the term ‘management representative’.
Few companies would have this as a job title but it implies an important
38
Perhaps this is not entirely true. The standard requires, for example, internal audits to be
carried out and documented. Internal audits would not necessarily have been carried out if
the system had not been implemented so there would be no records of them.
One task for senior management, when the system is being developed, is
to agree a quality policy and set quality objectives. The quality policy is a
matter for the organization and should define what it hopes to achieve.
A possible policy statement for a market research company (many others
would also be valid) might be:
Resource management
Another important task for management is to ensure adequate resources
are made available to implement the quality management system and
more generally ‘enhance customer satisfaction by meeting customer
requirements’ (Clause 6). Resource classes specified are human,
infrastructure and the work environment.
Some further design work may come after commission of the project,
such as questionnaire development, details of sampling and analysis
specification but these are subsidiary and follow the main design work
set out in the proposal. Separate steps in developing proposals may be
analytically distinct but they are often carried out more or less together.
These steps include:
Proposals do not prepare themselves, and staff must be selected for the
particular job and between them agree the research design, draft the
document and probably review and check it before delivery. Typically in a
smaller company, for an average type of project, the proposal and all its
design steps is the responsibility of one member of staff alone; generally
a senior researcher with the required technical skills. Often there is no
second person review (no one else may be qualified to review the
proposal). In larger companies and for major projects a number of people
may be involved, including a second review of the proposal document,
but the process is not different in essence. In both small and larger
companies, staff outside the ‘research executive’ team may have inputs
into preparing proposals and especially the data collection department,
but more often to do with timing and costing than research design. For
‘post contract’ design work – e.g. questionnaire drafting – the same
members of staff that prepared the proposal (or direct colleagues) are
often responsible, possibly with second person review (having the field
and possibly data processing managers look at a questionnaire before use
is good practice) but not in all cases. Questionnaires may also be formally
tested but full pilot research is usually carried out only if agreed as part
of the contract. The design process for continuous research projects and
multi-client work is often no different to that for ad hoc work. However,
there may be more time available, the design stages may be more
distinct and formal testing and piloting more common. The client may
even pay for the design stage of the project.
Clause 7 is not only about design; 7.1 defines requirements for ‘Planning
of product realization’ but these will normally be met by general
procedures for the research process, the research proposal or quotation,
such as data collection and data processing ‘work’ instructions.
Determination of requirements (Subclause 7.2) and especially the client’s
requirements are also normally covered by the proposal document or the
specification provided by the client.
A research company may carry out all processes ‘in-house’ or some may
be subcontracted, such as to fieldwork specialists. In that case,
BS EN ISO 9001 requirements are specified as Subclause 7.4, ‘Purchasing’.
Although this Clause applies to all types of purchasing, it is implicit that
bought-in inputs with a direct impact on meeting client requirements are
the main issue here. Suppliers are required to be controlled through, for
example, approved supplier lists and their performance is to be
evaluated. The product or service to be bought needs to be specified as
well as the quality assurance processes of the supplier and this may
favour suppliers which themselves meet BS EN ISO 9001 (this is implicit
rather than explicit in the standard). There is also a requirement to
ensure the product or service supplied by a subcontractor meets
requirements. In the case of physical products this may be by inspection
of the delivered product but in a service such as market research data
collection physical inspection is less feasible or effective. Areas such as the
number of completed interviews and whether the quota plan has been
set can be checked, whether interviewers carried out the interview as
required and appropriately will be less evident and assurance by
interview validation may be required; these could be specified in the
contract with the subcontractor.
The processes of the organization and the output (products) are required
to be monitored and measured (Subclause 8.2.3/4). Monitoring of market
research processes, in practice, will be limited. Internal auditing will
establish that procedures for the research processes are followed and
should identify whether these fail to achieve the intended results.
Examples could include bottlenecks, which in the context of market
research could be getting completed questionnaires back from
interviewers, manual data entry, etc., that tend to affect the agreed
timetable of projects or lead to a company failing to be able to offer the
turnarounds clients require. Inspection of the finished outputs of
research, for example, reports, tables, etc., can check conformity to
standard layouts, table presentation, grammar and spelling, but not
whether the research findings are basically ‘true’ (although professional
judgement may do so to some extent). These types of checks are likely to
be covered in quality management system procedures, as are checks
earlier in the research process, from proposal drafting onwards.
41
This is also referred to as ’third party assessment’ with self-certification being ’first party
assessment’. ’Second party assessment’ is by the customer and may be used to ensure an
uncertified supplier follows at least the key relevant requirements of a standard – e.g.
checking a fieldwork subcontractor carries out validation to the required level. The term
’registration’ is also used more or less interchangeably with ’certification’.
42
A certification body that is not accredited may well carry-out assessment to a high
standard but how can we be confident of this?
not just a one-off test; at least annually, the external assessors require
evidence of continuing conformity and to identify areas for
improvement.
This and the next chapter are about formal quality standards for market
research and the assessment or inspections that enable companies to
demonstrate compliance to the standards. The next chapter deals with
the two international standards that are relevant for market, social and
opinion research: BS ISO 20252:2012, Market, opinion, social research –
vocabulary and service requirements (the main research standard) and
BS ISO 26362:2009, Access panels in market, opinion and social research –
vocabulary and service requirements, which is specifically for access
panels are a tool used for recruiting respondents and collecting research
data. Whereas these ISOs for research are relatively recent developments,
the research sector has a long history of tackling quality, and this chapter
is concerned with the much longer established process standards and
associations which were the precursors to the international standard for
market research – the Interviewer Quality Control Scheme (IQCS) and
other national standards. UK initiatives, particularly IQCS, acted as
catalysts for international change and so we will use the history of the
development of IQCS and national standards to set the context for the
more recent international developments that are discussed in the next
chapter.
Interviewing problems
The majority of market research is based on data obtained by some form
of primary data collection and whether or not this work is done well or
badly has a very direct bearing on quality, on the ‘rightness’ of the
Interviewing problems
Obviously this structure was not conducive to the best quality fieldwork
since research companies had almost no direct control over interviewers
and could only hope that the work was done to an acceptable standard,
whatever that might be. Interviewers also had little reason to have any
feeling of loyalty to research companies since their ‘employer’ was a local
supervisor who tended to have more regard for their own team than the
‘clients’, the research company. In any dispute about quality of the
output the supervisor, as likely as not, was defensive about ‘their team’.
The tangible rewards offered to interviewers were low – unskilled or
semi-skilled rates just above minimum wage in some cases – but
supervisors (who tended to be drawn from the pool of experienced
interviewers) did a bit better. Interviewing was a part-time and spasmodic
job and was generally regarded as offering ‘pin money’ rather than a
living and, it was assumed, well suited returners-to-work (generally
women with families) and pensioners with a bit of time on their hands.
43
According to the AA/Warc online Expenditure Report, total UK advertising revenue was
£15.5bn in 2010 compared to the £2.06bn total research revenue reported in the 2010
annual survey produced by The Market Research Society.
Also, due to their relatively small size, many research companies lacked
much in the way of in-house administration and were not, therefore,
able to manage interviewers directly and were more than happy to turn
over the responsibility to a local supervisor. The low rates offered to
interviewers reflected supply and demand; with little or no training
required to become an interviewer, supervisors could always find
someone to take on work and the pay levels fed into the prices charged
by research companies to clients; research companies could not afford to
pay more than the going rate. It should also be remembered that
historically research did not have much competition from other sectors
offering more favourable work terms. Before the era of working time
regulations, and flexible and part-time working, research interviewing
was one of the few jobs where you could work your own hours, could
accept or reject an offer of work and where you needed little or no
qualifications to get into it.
There have also been other developments which have favoured closer
contact and a degree of control between the research company and the
interviewer. One factor is that most interviewers are no longer generally
classed as self-employed. European legislation such as the Working Time
Directive has resulted in a reclassification of most interviewers. In the UK,
for example, most interviewers are classed as ‘workers’ with the
exception of qualitative recruiters (specialist interviewers) who still tend
to be self-employed. This is in recognition that employers have a
relationship with the interviewers, some continuity of contact and
therefore some obligations to them. Although most interviewers are not
legally classified as ‘employees’, except in cases where research companies
have actively decided to employ them (this is mainly found in telephone
Interviewing problems
Interviewing problems
are no longer typical of the population due to being more familiar with
research (we’ll discuss more about panels when we get to
BS ISO 26362:2009 in the next chapter).
The problem with cheating of any sort is that it is corrupting. Once done
and if not detected, it will be seen as easy and will almost certainly be
repeated. What started off as a one-off to solve a pressing problem
becomes standard practice. Research companies’ remedies against
cheating include dissuasion through morale building and training, close
contact with interviewers, supervision where possible and validation
methods to detect it.
The above puts all the blame on interviewers for both bad practice and
cheating. But this is hardly fair. Research companies can do much to help
their field-workers through better training and administration.
Furthermore, researchers must shoulder some of the blame. Poor
instructions, badly drafted questionnaires, quotas that are virtually
impossible to meet, and other design defects make the interviewer’s task
that much harder and are no encouragement to good work. Part of the
problem is that research executives often have little or no experience of
the fieldwork they are ordering. In an ideal world, all trainee research
executives should serve at least six months as interviewers.
44
In most countries legal requirements means that monitoring can only carried out if the
interviewer is aware this may occur. They do not need to know when exactly it would
occur (this would defeat the purpose) but that it may. Most research companies cover this
off legally by stating in their worker contracts, staff handbooks, etc. that monitoring
and/or recording of interviewers may take place.
Training can also be carried out more frequently and at lower cost than
for field interviewers and every project can be briefed adequately, with
problems solved as they arise. Morale is also often stronger because
interviewers work together and at the research company. Their pay and
conditions may be no better than interviewers out in the field (many
interviewers are still paid by the hour) but with only a modest
management effort they can develop good team loyalty and commitment
to the company. Unlike field interviewers, they often work for one
research company at a time and perhaps for long periods (although every
call centre will have a percentage of the staff that is temporary). Due to
its nature, telephone interviewer quality is, therefore, much easier to
control and effective measures can be taken by research companies at
modest cost (which is less true in the case of field interviewers). However,
complacency should be guarded against and the effects of telephone
interviewer variability cannot be assumed to be minor.
As well as normal field and telephone interviewing, there are some more
unusual forms of data collection which have specialized quality concerns.
These include the various sorts of observation techniques associated with
retail auditing and the increasingly used ‘mystery shopping’ approach.
With the latter, one school of thought believes that any more than
limited training actually reduces quality. It is the reactions of ‘naïve’
shoppers which are sought.
The IQCS has two main functions: defining and revising minimum
standards (in the light of feedback from members), and administering an
assessment or inspection scheme to establish members’ and candidate
members’ compliance with the standards.
Interviewer appraisal
The second part of IQCS standards – appraisal – can be regarded as a
continuation of initial training. The latter may be deemed adequate for
the interviewer to start working but what is to ensure that he or she
does not forget what was taught or slip bad habits? Field work, at least,
is carried out without any direct and on-going supervision. Frequency and
methods of appraisal are specified in IQCS standards. As a minimum,
every field interviewer used on five or more projects/assignments a year
must be appraised annually and this must include an element of
Interviewer validation
Validation is principally about checking interviewers’ work directly with
the respondents concerned. In telephone units this can be done by
monitoring – listening-in without either interviewer or respondent
knowing that this has been done. Such validation would certainly catch
cheating but, as argued, this is unlikely to be a major problem in a
telephone room. Telephone monitoring is, however, an effective method
of establishing levels of interviewer variability and reducing it through
follow-on training and by feedback to the interviewers.
The above summarizes the requirements of the three key parts of the
standard, all of which are also requirements of the data collection
element of BS ISO 20252. In addition there are important requirements
for adequate record keeping and documentation. These enable the
research company to control fieldwork quality management and in
addition are the means by which IQCS, through its team of inspectors,
establishes member or candidate member compliance with the standards.
Which records must be kept are specified but the general principles are
that project records have to show which interviewers worked on the job,
what they did and what validation was carried out. In addition
interviewer files must be maintained to show the training and appraisal
records of interviewers, the work they have been assigned, the outcome
of validation of their work and various administrative details. The
standards also specify documentation that must be given to interviewers
including training manuals and an annually issued MRS identity card.
What is not specified in IQCS is any need for formal quality assurance;
systems to assist research companies in ensuring consistent compliance
with the standard.
IQCS standards also cover other areas which we have chosen not to
discuss at all. The most important of these are the requirements
concerning field based supervisors. It is they who largely recruit, train and
appraise interviewers.
Inspections are carried out by a small team employed, part time, by the
IQCS. The individuals concerned all have practical and long-term
experience of market research fieldwork. Inspections are carried out
when a company applies to join the scheme and annually thereafter. The
inspection work takes place at the research companies’ office with the
time spent linked to the size of the company (effectively a minimum of
one day) and very largely consists of checking records with cross
referencing of project and interviewer files. This establishes whether
training, appraisals and validation have been carried out to the
standards, although, strictly speaking, the inspection only proves that
records exist indicating that the requirements have been met. There is,
therefore, some potential for the companies to cheat through false
recording but this probably happens to only a minor extent. It is hard to
imagine that a company would go to all the trouble involved without
having a basic willingness to comply. The IQCS has been criticized because
its inspection is largely concerned with historical records rather than
directly establishing fieldwork competence. Realistically, however, it is
hard to envisage how any inspection or assessment, at a realistic cost, can
be anything other than largely record based. It should also be said that
inspections are no walk-over. Every year, some candidate companies fail,
as do, occasionally, existing members.
Is it all enough?
One point to make immediately in considering whether IQCS standards
are adequate is that they must be acceptable, taking into account the
costs entailed, to the companies required to implement them. It is easy to
think up more and more quality hurdles, but unless members are willing
to follow them they will vote with their feet, and leave the scheme.
Similarly there is no point in being critical because IQCS standards do not
address the full scope of data collection – there are other standards for
this (discussed in the next chapter). This is not the intention; only the
work of interviewers is addressed and not the important contribution of
design work (a bad questionnaire will fail to produce valid data
regardless of the interviewers) or the project management of fieldwork.
After a rather bumpy start – with some large research companies publicly
castigating the standards as having no value – the national standards
gained traction in the marketplace and following their launch became an
accepted feature within the research sector in the same way that the
IQCS had. Following the success of the national standards for the whole
research process, and the increasing globalization of the research sector,
there was some demand to develop one international standard for
research. The national standards – particularly the UK’s BS 7911:2003,
Specification for organizations conducting market research, which had
influenced other national standards such as Spain’s and Australia’s –
became the foundation for the first international standard for market
research – BS ISO 20252 – originally issued in 2006 and subsequently
updated in 2012. This in turn encouraged the development of a further
45
From the introduction to BS ISO 20252:2012, Market, opinion and social research –
vocabulary and service requirements.
As with the national standards which preceded it, BS ISO 20252 was
developed to cover the whole of the research process and draws heavily
from various national standards that had already been developed,
including the UK standard BS 7911:2003, Specification for organizations
conducting market research. Following the creation of the international
standard, most countries withdrew their national equivalent in
preference for the ISO. However, there are still many national standards
bodies which have yet to implement BS ISO 20252 and the reasons for
this and the impact this has will be discussed later in the chapter.
By and large what is not covered is any attempt to define what is right
and/or valid research design. When and where qualitative rather than
quantitative research should be used, sampling and questionnaire design
principles, on what basis and how data can be generalized and how
inferences are drawn, are all examples of what is not covered in this
standard. These all touch on issues of quality as excellence rather than
relative quality. It is not that these areas are considered unimportant to
quality in research. On the contrary, these are fundamental matters which
would be inappropriate to be dealt with within ISO type standards and
assessment frameworks. Instead issues of what constitutes ‘right’ research
are professional issues to be defined by an informal consensus, as with all
matters relating to a profession. At the leading edge of research practice
there may be debate and dissension on what is ‘right’ but nevertheless,
there is a large body of research practice which is reasonably certain,
accepted and understood by all practitioners of any experience.
The reason that ISO could not sensibly define right methods is because of
the difficulty in reaching a consensus on what is right. This is primarily
because of the inherent diversity and complexity of the business
problems that research seeks to address. Every business problem requires
a bespoke research solution, some of which may be relatively
straightforward but others much more complex. It is these ‘softer
elements’ of quality research design and problem definition (Smith, D
and Dexter, A 1994) that means that research ISOs and other research
quality standards such as IQCS will not be suitable to resolve these types
of research issues, whilst still being fundamental to ensuring robust
research processes. This in turn reflects the shifting intellectual ground on
which market research is built and that market research is still a very
open and evolving profession. Any attempt to define ‘right’ research, in
simple statements, would inevitably have been criticized by people who
could reasonably claim authority within the sector. Whatever else, ISO
research standards development requires wide consensus which would
have been impossible to achieve for professional service standards of this
nature.
There are, therefore, good practical reasons for the course taken but it
should be noted that in other knowledge businesses, ‘right’ methods are
defined and set as standards. Accountancy is a good example. There can
be legitimate debate on how goodwill should be treated in accounting
statements, but this has not prevented standards from being set.
However, accountancy is not market research. Probably researchers would
contend that although some aspects of accountancy might be
sophisticated and require extensive knowledge, because it deals with
normative knowledge, it is all more certain than what is appropriate in
researching markets and consumers. No doubt accountants would not
wholly concur with this view. However, the fact remains that
standardizing ‘right practice’ in market research has not proved to be
practical as yet and is unlikely to be so in the foreseeable future.
Ultimately excellence in research comes down to a combination of
suitable quality standards (such as ISOs), appropriate professional ethical
codes and individual knowledge and competence. Only when all these
aspects are applied properly can claims of research excellence be made.
To blame ISOs for poor research is a bit like criticizing a footballer for
being bad at cricket; it just misses the point and purpose of ISOs.
• Meeting requirements,
• The execution of a research design,
• The services supplied to a client over and above ‘right’ research
practice.
While the standard does not seek to address what is ‘right’ research
design, it does cover how research design is put into practice, with
various controls and checks specified to increase the likelihood that what
is intended in the design happens in practice (and at the first attempt).
This, then, is the general nature of the BS ISO 20252 standard. We shall
now consider the contents of the specific sections of the published
standard in the order in which they are set out in the document.
In the 2006 version of the standard this Clause was titled ‘quality
assurance’ and the change is of wording only and has been made solely
to comply with ISO drafting rules;46 the section is still concerned with
quality assurance or management and quality systems. Since the original
quality research standards such as BS 7911:2003 and IQCS were born out
of the perceived limitations for research of the early versions of the
broader quality assurance standards, such as the BS 5750 and
BS EN ISO 9000 series, it may seem odd that the first substantial section
of the ISO is in effect devoted to quality assurance. However, the need
for companies to ensure that they had effective quality assurance had
been accepted from the start. In this respect it was intended that
BS ISO 20252 should overcome a weakness of standards such as IQCS;
which do not contain requirements for internal quality assurance
methods.
46
ISO policy is to limit standards for quality management to broad, all-business standards
such as BS EN ISO 9001 and to not allow industry-specific quality assurance standards, as in
the original BS ISO 20252 produced in 2006. As a compromise with the drafting committee
it was agreed that the section could remain if ’quality system’ was renamed as ’research
process management system’ – old wine in a new bottle.
47
BS ISO 20252:2012 has a clause relating to confidentiality of research (3.2). This is not an
explicit requirement of BS EN ISO 9001 though it is implicit in various ways. Also
BS ISO 20252:2012 requirements for subcontracting are more specific to research activities,
but arguably no more demanding than in BS EN ISO 9001.
Eastern Europe, data analysis in the UK, etc. The reasons for this are
primarily financial, in that certain countries can provide certain elements
of the research process significantly cheaper than domestic operators
(and often these services will be provided by companies that are part of
the same group as the main research service provider). However, this cost
saving approach does not come without its risks.
The scope of this module of the standards covers the tasks typically
carried out by a research ‘executive’, including responding to research
requests, contents of research proposals, research design and proposal
writing, questionnaire development, briefing of interviewers and
reporting of research results, as well as overall project management and
liaison with the client. However, the ‘executive’ can go by any title and
the standards allow for more than one person to have executive
responsibility for a project. There is, therefore, sufficient flexibility to fit
any organizational structure for both ad hoc and continuous research. At
some critical stages in the process – e.g. the sampling procedure,
questionnaire or discussion guide development, data collection, analysis,
etc. – ‘monitoring’ is required; this could include checking the executive’s
work to ensure that such activities are conducted according to agreed
The requirements of the executive standard fit into the quality assurance
requirements of Clause 3 of BS ISO 20252 and those of BS EN ISO 9001. In
terms of the latter, the standards specified relate mainly to product
realization, customer-related processes, design and development and
production and service provision.
Data collection
The data collection section is the most extensive in BS ISO 20252 but this
is understandable as arguably the research data collection stage is the
most prone to error. The already well-established UK interviewer
standards from BS 7911:2003 (which were originally based on the IQCS
standards) were the start-point for the data collection requirements
detailed in BS ISO 20252. Following the completion of the first edition of
BS ISO 20252, in 2006, the IQCS amended its standards so the two would
remain compatible. The principal data collection requirements of
BS ISO 20252 covering interviewer activities (training, conducting
fieldwork, validation methods, back-checking etc.) are already detailed in
Chapter 8. The additional elements of Clause 5 are mainly concerned
with self-completion data collection, observational data collection and
data collection from secondary sources and the revised 2012 version
includes these new additional areas compared with BS ISO 20252:2006.
Standards for carrying out some of the main steps in data processing are
prescribed and these include developing a code frame for open-ended
responses, coding training, data editing, data analysis, and the
presentation of output including composition of tables and electronic
data delivery.
Assessment
Assessment and certification to ISO standards may be carried out by
independent bodies (‘third party’ assessment).48 In some countries, such
as the UK, the assessment companies follow a national assessment and
certification specification and are also separately accredited (in the UK by
the United Kingdom Accreditation Service) – see also Chapter 7 for the
need for such accreditation to ensure that they are suitable qualified to
carry out assessment for the specific ISO standard.
During initial assessments all parts of a research company and all relevant
parts of the standard are covered. The work involved includes both
checking the quality system documentation to ensure the requirements
of the standard are fully met and documented and then establishing
whether the system is actually followed in practice. These are separate
steps in the process and the review of documentation may be done
‘offsite’ with any issues arising and the need for changes (e.g. some
requirements not adequately addressed) discussed with the research
service provider before starting the second stage. The length of time
involved in initial assessments depends very much on the size of the
research company concerned but is unlikely to involve less than a full
day’s work by the assessment body’s staff. A greater input will be
involved for larger companies. Where ISO standard assessment is
combined (e.g. BS ISO 20252 and BS EN ISO 9001), the actual work
undertaken by assessors is not likely to be significantly different to an
BS ISO 20252-only assessment. In either case the same quality system is
assessed.
Companies cannot pick and choose which parts of the standard to follow;
all requirements defined in the whole standard must be met if the
relevant process is carried out at all. All research companies are required
to meet the quality assurance requirements of the BS ISO 20252 standard
and assessment of this part is either by the assessment body against the
quality assurance model defined in Clause 3 of the BS ISO 20252 standard
48
BS ISO 20252:2012, as all similar standards, does not specify how assessment should be
carried out. There is a general principle that all such standards can be assessed by a first,
second or third party. First party is internal self-assessment, second is by the client and
third is by an independent party.
Implementation
Implementation of BS ISO 20252 is no different in principle (or largely in
practice) to BS EN ISO 9001. In either case research companies need to
develop a unique quality system which both matches the requirements of
the standard and the particular needs and structure of the business. Once
the system is designed, it is then a matter of getting the company and its
staff to work to the system. When the system is working, assessment can
be arranged. The UK assessment specification requires that the system
must have been operating for at least three months before assessment
(in other countries the period may be different), although in practice the
elapsed period is likely to be longer; to carry out a meaningful
assessment enough time needs to have elapsed for there to be adequate
evidence of compliance or not.
Linked to this is the fact that research companies working to the same
minimum standards will be competing in a more comparable way.
Research prices can always be brought down if corners are cut, but with
research companies working to the same minimum standards, clients are
protected from suppliers tempted to win business on the basis of offering
research which seems to meet the need but is in fact deficient. Research
companies implementing BS ISO 20252 will certainly incur costs, as do
those which are members of IQCS, and they will wish buyers of research
will be able to distinguish suppliers of this standard from others which
may offer lower prices. If quality is compromised, lower prices do not
mean better value. Poor research is never a bargain.
One of the main difficulties with BS ISO 20252 is its fragmented global
implementation. Most countries which had domestic research standards
were able to establish an ISO assessment and certification scheme
49
From the Introduction of BS ISO 26362:2009, Access panels in market, opinion and social,
research – vocabulary and service requirements.
The US was one of the countries which led the way in developing online
panels, Inside Research (2009) estimated the total US spend on online
research in 2009 at about $2 billion, the vast majority of which was
supported by online panels. In ESOMAR’s Global Market Research Report
2011, online quantitative research methods accounts for 22 per cent of
research spend, a significant amount of which will be via access panels.
Increasingly there has been concern about panels, particularly in terms of
panel quality. Procter & Gamble, and a number of other big research
buyers, reported that various online research panel projects appeared to
produce inconsistent results. Questions began to be asked about whether
panel respondents were in fact professional respondents, whether
frequently used panel members became increasingly unrepresentative of
the population, whether some respondents were taking part in too many
research projects (including amending screener questions to qualify for
participation) and the rapidity with which some respondents were
completing research surveys and whether such results could be relied
upon.
To stem this flow the sector needed to take action. The sector’s response
needed to be a layered approach due to the nature of the problems,
which were a mixture of research process, design and buyer
education/awareness. Initiatives such as the American Association for
Public Opinion Research Opt-In Online Panel Taskforce was formed in
2008 to review the ‘current empirical findings related to online panels
utilized for data collection’ with an aim to ‘provide key information and
recommendations about whether and when opt-in panels might be best
utilized and how best to judge their quality’ (AAPOR, March 2010).
Similarly ESOMAR developed its 26 Questions to Help Research Buyers of
Online Samples with the aim of aiding researchers and their buyers to
decide whether an online sampling approach is fit for purpose in relation
to specific research objectives. Other countries such as Australia
developed domestic process standards for online panels (the QSOAP Gold
for online access panels). Other initiatives such as the Association
Collaborative Effort (‘ACE’) a collaboration between six research
associations (primarily in the US) was formed to address and improve
online research quality, and this has resulted in initiatives such as the
release of a set of online research definitions in 2009. New initiatives
continue to emerge, in 2011 ESOMAR and CASRO50 announced plans to
50
CASRO is the US Council of American Survey Research Organizations.
Requirements
The requirements of BS ISO 26362 are based on the principle that any
research access panel will be ‘actively managed’ rather than just a
database of respondents who have at some time in the past indicated a
willingness to participate in research. The panel also requires panel
providers to have a quality management system in place which covers all
the requirements of BS ISO 26362, plus the organization and
51
From the Scope section of BS ISO 26362:2009, Access panels in market, opinion and social
research – Vocabulary and service requirements.
What is required?
What is required of staff to deliver quality? There are two major and
linked requirements: commitment and skill. Commitment is about
accepting the importance of quality and wanting to deliver it, all the
time. Various morale building approaches – management leadership
(including by example and exhortation) – all have their part to play, as
does staff involvement in quality improvement through internal audit
and quality teams, special task groups, and involvement in the
problem-solving aspects of a quality system. By these and other means,
the managers and staff, at all levels in a company, can develop a regard
for quality – the need to strive for excellence in research, meeting or
more than meeting clients’ needs, working to the specification, and
maximizing efficiency. However, for any of this to work, staff need the
appropriate knowledge and skills to perform their jobs in the best
possible ways. Regardless of the morale level, quality cannot be delivered
if the knowledge and skills are not there and in the long run frustration
will undermine morale.
To meet a short-term need, some knowledge and skills can be bought in.
Staff can be hired to enable a company to offer a new sort of expertise
and sometimes this is essential. If a quantitative agency decides to extend
its services to qualitative research, there is no alternative but to recruit at
least one or two key people if they want to do it properly. To fill a
shorter-term need, temporary or freelance staff can be brought in.
However, recruitment can never be the only or even the main means of
building up a knowledge and skills-base. Existing staff have to be trained
and their skills developed. This process has to be continuous because
knowledge moves on relentlessly; no one ever knows enough or has
enough applied skills. Even when new staff members are taken on, they
must also receive some in-house induction training to fit them into their
new team, corporate processes, etc. No company, least of all a market
research company, can ever ignore learning and staff development.52
52
We shall not make a hard distinction between staff training and professional development.
However, training can be thought of as short-term and structured around specific learning
outcomes, while professional development is more of a personal development plan based
around specific, identified learning needs which may need to be fulfilled via a variety of
formal and informal development activities.
What is required?
In case the reader is in any doubt, the above is the bad model of
training. It may be an exaggeration. Perhaps market research companies
are not that bad! But in many cases the organization of training is
probably not much better than this. So what is a better model?
Assessing skills
Once the professional development needs are defined, the next step is a
‘learning and skills’ audit. To what extent are the required skills present
already and what gaps exist? Such gaps can be both quantitative
(numbers of staff with the appropriate skills) and qualitative (key skills
entirely absent or at too low a level). Thought can then be given to how
the gap is to be filled. The two broad solutions are buying in –
recruitment – or internal training. Qualitative gaps are more likely to
involve the former, although there are no hard and fast rules. The
professional development plan (but not the broader business plan),
however, will be less concerned with the recruitment side.
With needs at both company and individual level defined, an action plan
can be drawn up covering the forthcoming year and possibly longer
term. The plan will set priorities in terms of what training is to be given,
to whom, and with what amount of the budget (there may obviously
have to be compromise and trade-off at this point). The methods of
delivering learning also need to be decided. This may involve sending
staff on external training courses or developing a bespoke course to be
delivered exclusively for staff in-house. However, it is a mistake to assume
that effective training always has to involve external, paid-for courses. In
any company, there is always potential for skill and knowledge sharing;
senior staff and staff who have been on external courses can hold their
own seminars. There is also the development side: skills and knowledge
can be self-taught and, in truth, any theoretical subject can often be as
well mastered by old-fashioned reading as by attending seminars. With
the advent of online resource there is a wealth of information that is
now readily available as either free training or low cost online training.
Informal coaching, hands-on learning and buddy systems (where staff are
put with more experienced staff from whom they can watch, listen and
learn) are also effective means of knowledge and skill delivery and the
only drawback to them is that they are often not recognized for what
they are. Staff may be well trained, but might not think of informal
What is required?
methods as training; often the price tag determines the value that staff
place upon it but often the best training can be the free hands-on
learning from colleagues.
Evaluating skills
To complete the training cycle, evaluation of the implemented plan is
required, with any lessons learned applied to the next year’s plan. First of
all, it should be established whether the plan was substantially delivered.
Perhaps for one reason or another some of the training was just never
delivered. Why? There may have been good reasons and these should be
considered next time round. Also the plan may have had to be changed
to meet circumstances – the need for additional knowledge and skills
may have been recognized over the year. Planning is essential but it must
not be so rigid that it cannot adapt as needs change.
One of the most comprehensive NOS which covers market research are
those developed in the UK and we will describe these to illustrate how
NOS can be used for developing and assessing research staff quality and
competence.
The research NOS are part of a wider set of NOS covering the whole of
marketing and marketing communications. The basis for marketing NOS
is a functional map (see Figure 9) which in addition to setting out the
specific marketing and marketing-related activities, also includes more
general knowledge and skills such as management and finance (a
comparable and mirror set of standards for social marketing including
social research has also been developed in the UK).
Within each of the eight elements of the functional map there are a
multitude of units (around 100 of them) each describing a part of the
function and identifying performance criteria, behaviours and knowledge
and understanding, which individuals involved in that activity should
display. An example of one of these units – for collecting market research
data is shown overleaf. As you can see there is nothing revolutionary
here; someone involved in collecting market research data should be able
to select the right research environment, use the research techniques,
have good communications skills, and understand the legal environment,
etc. The value of NOS is that they are a ready-made toolkit of some key
standard functions that exist within marketing and market research. In
the UK the NOS are used to develop vocational qualifications, i.e. any
government accredited qualifications must clearly map against the
relevant NOS. On a more practical day-to-day basis, the NOS can be used
by employers to assess and benchmark staff performance against a set of
established criteria, help to identify skills and knowledge gaps that need
to be filled and can be used in appraisal schemes and developing job
roles. They also provide managers with a tool for workforce management
and quality control, can assist with setting individual and team objectives,
and can assist with designing and developing in-house training
programmes (see Figure 10 for an example of how it can be applied in
practice).
Investors in People
Investors in People
The IiP standard was developed in 1990 by a group of employees and
employers, and was launched nationally in the UK in 1991. In October
1993, Investors in People UK was established as the body responsible for
the standard. Since its launch IiP has been adopted in countries
worldwide in Africa, Europe, North America, the Middle East and
Australasia (see Investors in People International for more details); for the
purposes of this chapter the UK publication, as the originator of IiP, will
be used.
The standard has become more flexible, with businesses now able to
concentrate on those parts of the IiP framework which are most relevant
to its business.
53
Bourne, M and Franco-Santos, M; Investors in People, Managerial Capabilities and
Performance – A Study by the Centre for Business Performance, Cranfield School of
Management (2010).
54
Institute for Employment Studies, Does IIP add value to business? IES Report (2008).
55
Based upon organizations listed in MRS publication, the Research Buyers Guide 2011 and
accredited with IiP.
Investors in People
prescriptive about the form this takes. IiP focuses on outcomes; what
needs to be achieved rather than how to achieve it. Commitment also
implies some longer-term vision or purpose of where the company is
going and it is essential that this is effectively communicated to all staff
(including the contribution they will make to realizing it). The vision, of
course, does not have to be overambitious; it is perfectly satisfactory if a
company wishes to stay the same size. Commitment to growth is not
essential (indeed can be fatal if an organization falls into the trap of over
trading) and staying in business has to be core to any business ambition.
The indicators related to this part of the standard, such as mission
statements, core values, business strategies, key performance indicators,
etc., cover visible evidence of the commitment with staff involvement in
development. When it comes to assessment, what staff know, believe and
feel is as important as any documentation.
As a document the IiP standard is relatively short and user friendly. There
are no difficult terms and the language is easy and non-technical. There
is no need for any sort of translation to relate the standard to the
practices and structure of any particular industry. A fair criticism of the IiP
standard, however, is that it can appear to repeat itself in several places,
but this is quite easily untangled.
IiP is non-prescriptive and very flexible. However, IiP is not in any sense
an alternative to quality assurance systems. It addresses quality but from
a different and complementary approach. There is no requirement for an
BS EN ISO 9001- or BS ISO 20252-accredited company to seek IiP or vice
versa but they hang together well. If a formal quality system, such as
those found in the ISO standards, is in place, it is sensible to integrate the
documentation and records of learning and development into the formal
system. If the documents and records are needed they should be well
kept and a formal quality system is, among other things, an effective way
of ensuring this.
56
Investors in people, http://www.investorsinpeople.co.uk/interactive/Pages/default.aspx
[accessed 12 April 2012].
Assessment to IiP
Assessment to IiP
Assessment is carried out by independent consultants from local Investors
in People Centres. Normally the process will begin with a meeting to
discuss the business priorities, what is involved in assessment and to
agree how the process will be conducted. An important part of this
discussion is agreeing the scope of the assessment and discussing an
organization’s business priorities, performance targets, etc.
The on-site assessment may take several days for a medium-sized research
company and rather longer for the largest or those with several sites.
Small companies may be completed in a day or two. Most of the time
during the assessment is taken up with discussions and meetings and
involves as many staff as possible. IiP assessments are entirely
interview-based. The paperwork can provide practical examples and
evidence, but the assessment will be based on the discussions with staff.
IiP: In summary
Being a recognized Investor has some kudos and may impress clients or
even potential staff to some extent. However, it is very doubtful that
work will be gained just on the strength of the plaque. If IiP is worth
having, it must be because the processes necessary to meet the standard
produce benefits quite unrelated to assessment or recognition. If a
company is following anything approaching the suggested ‘better model’
of professional development it will be doing nearly everything needed to
gain IiP recognition. And the commitment to learning and development,
implicit in the model, is in any case essential to delivering quality and
pursuing business success. Therefore, why not seek IiP recognition? There
will be some assessment costs but the status that is gained – and this
includes internally as well as externally – will be at least commensurate.
Indeed, based on the evidence of numerous research studies into the
impact and effectiveness of IiP (e.g. Bourne and Franco-Santos, 2010),
there is good evidence to suggest that organizations that have IiP
perform better financially and thus any investment in the cost of IiP is
more than recouped in financial growth.57 In addition, the process of
seeking IiP is itself a spur to action. Learning and development are of
57
Of course it is possible that the companies that perform better are more likely to be
receptive to investing in IiP.
IiP: In summary
course essential but the prospect of assessment gives that extra reason
for doing things right and doing them when they should be done.
The first is time available for the role. This will vary with the size of
company and the extent to which that expertise is bought-in. The role
will not need to be full-time, except in a large company, but some diary
space is essential, including in the busy business periods. Whoever takes
on the role needs the backing of the quality champion and other senior
staff to make time available. The project leader does not have to be a
hands-on professional researcher but does need to understand, in at least
broad outline, what is involved in the research processes. He or she also
needs to know the structure of the company and who does what; in a
smaller company this is not likely to be a problem. Related to knowing
the company are people-skills – in communicating, persuading, and
obtaining information from colleagues (this should not be too hard if the
project leader is a researcher!).
Finally, there is the need for specific expertise about quality systems and
BS ISO 20252. Sometimes this may have been gained in another company
but this would be fortuitous. The necessary knowledge can come from
various sources such as courses (in the UK, relevant courses and
workshops are offered from time to time) and reading…and not least
If the right consultant is brought in, not only will expertise be provided
but some of the internal workload of the project can be reduced,
including drafting quality system documentation. Furthermore, with,
hopefully, practical previous experience as well as expertise and possibly
less distraction from day-to-day work, a consultant will enable the whole
project to be completed more quickly than by someone learning as they
go along. Also, the consultant may reinforce the commitment; he or she
is costing the company money and the benefits need to be reaped.
However, there are drawbacks to using consultants (apart from costs),
and the biggest is that the project and resulting quality system is
something brought from outside, bolted on and therefore in danger of
never really being ‘owned’ by the company itself. A poor consultant may
also try to impose standard solutions without taking account of how the
company functions. As will be discussed, the documented quality system
should be unique to a particular company and off-the-shelf manuals will
not work. The question of whether or not to use an external consultant
is, therefore, something to weigh up and debate. In most cases it is
perfectly possible to do it all in-house, accepting that the project may
take longer.
Apart from people and their knowledge, other resources needed, beyond
normal office facilities, are few. One to mention though, and it should be
obvious, is a copy of BS ISO 20252. The UK version of this publication,
BS ISO 20252, is published by the British Standards Institution (BSI). In
other countries the national standards body is likely to be the source. To
58
In the UK, accredited certification bodies are not allowed to act as consultants. They can
and do offer training related to quality systems and may be willing to recommend
individual consultants.
The cost of implementing and being assessed to BS ISO 20252 will vary
from company to company depending on size, level of internal expertise
and the degree of change in working practice required to meet the
standard. The main budget headings include consultants’ fees (if used),
the cost of internal staff time (salary and opportunity costs), any extra
process costs (e.g. interviewer validation) and the charges made by
certification bodies for assessment. Quotations can be sought from
consultants and certification bodies but internal costs need to be
estimated, usually after carrying out an initial audit.
The initial audit can be carried out by a consultant who will (and should)
be familiar with the standard and its requirements. He or she, however,
will not know the company and its staff and so will need to work closely
with the project leader. Let us assume, however, that the initial audit is
carried out by the internal project leader without external help; the first
task will be a close study of the text of BS ISO 20252 and an
understanding of its requirements. The project leader can also decide at
an early stage, perhaps after some discussion with other staff, which
clauses and requirements do not apply because the processes are not
relevant to the work of the company. It is then a matter of going
through all remaining clauses and parts of clauses and establishing
whether current working practice is or is not in line with the standard.
The requirements of the standard are set out in Clauses 3 to 7 and these
can be covered in order,59 in each case establishing whether the
requirement is currently being met (or not). For example, Subclause
4.1.3.8 of BS ISO 20252 requires proposals to describe whether individual
steps are to be subcontracted. Does the company subcontract any
processes and if so is this stated in proposals? However, a simple yes or
no response is often not adequate. Firstly, even small companies vary in
how a particular process is carried out between departments or between
individual members of staff. One research director, for example, may
mention subcontracting when drafting proposals whereas another may
not. Similarly, subcontracting of data collection may be included but data
entry subcontracting not. Also, a particular requirement may be
considered good practice and generally met but not in every project.
Clearly, therefore, in carrying out the initial audit, the project leader
needs to do more than sit at a desk and think; establishing current
practice requires discussions and questions across the company. We
suggest that conformity to each requirement is shown on a simple scale
of at least three points: the requirement is currently met in nearly all
cases, met in about half the cases and met rarely or not at all, plus,
where the requirement is not relevant to the processes carried out.
Obviously this needs recording e.g. ‘pencil’ notes on a copy of the
standard or on a specially designed checklist.
When all clauses of the standard have been adequately audited, the
result can be reviewed. Many of the requirements will be found to be
met wholly or at least generally. Some will not be met and changes in
existing working practice will be needed for conformity. What is involved
and the implications of making the necessary changes will vary. In some
cases the change will not be hard to make and entail no real costs; e.g.
Subclause 4.1 lists what should be included in proposals (where relevant
to the project) and adding some of these to the current format should
not be difficult. Other changes may involve significant extra costs. If
phone interviewing, for example, is carried out but without interviewer
validation to the level required (Subclause 5.4.3) then making this change
will have cost implications. The senior staff of the company must,
therefore, consider carefully the feasibility of making the required
changes and on this basis decide whether to go to the next stage or not.
For most companies the changes to be made will not be too daunting or
costly.
59
Clauses 4 to 7 roughly follow the logic of the research process from proposals through to
reporting. Clause 3, though, is more concerned with managing the quality system itself and
it is unlikely that many of these will be already met and they may be a bit more difficult to
understand.
In other words the system has to cover all requirements relevant to the
processes undertaken by the organization.
The documented system needs to include all the process steps relevant to
BS ISO 20252 requirements. Additional processes not covered by the
standard can also be usefully included in the quality system. The need to
send clients invoices, for example, may be sensibly included. However, as
we discuss shortly, making the system over-detailed and covering
additional areas produces its own problems.
For a company implementing BS ISO 20252 for the first time, the
documentation should describe current working practice except where
changes are required to meet BS ISO 20252 requirements, as identified in
the initial audit. The requirement that the system is ‘auditable’ means
that it should be possible to establish, by means of some sort of record,
that the system and specific parts of it are followed by the company after
the event e.g. if a client is required to ‘sign off’ a questionnaire, can we
establish at a later point that this has actually been done? The need for
an auditable system needs to be constantly considered when drafting
documents but there are dangers of ending up with a complex system of
records whose only purpose is to satisfy the quality system, with no other
practical use. Such a system will be a burden to all and in the end will be
a failure. Ideally, any record of following the system should also provide
practical benefits in management. Wherever possible the record should
also be part of documents that are produced as part of the research
processes themselves. A questionnaire, for example, is its own record;
there is no need for a record to show a questionnaire has been
produced.
Example of an introduction
Like all good documents, ARSP procedures have an introduction. It is less
than a page but is more than just formal significance; it covers some
requirements of BS ISO 20252.
60
From here on ‘procedures’ will be used for a documented quality system; other terms such
as quality manual can equally be used.
The final paragraph of the introduction relates to the need, set out in
BS ISO 20252:2012 (Subclause 3.3.2), for control of key documents:
not mentioned – this is also covered in Appendix 2). Some general points
about the procedures include:
In other procedures, other records are referenced and these also meet
the need for auditable procedures but in nearly all cases in procedures up
to 7, the records are either produced by the research process (e.g.
questionnaires) or serve a more general business need. Some records may
be no more than emails to colleagues, clients or subcontractors and these
are quite adequate providing they are easily locatable and retrievable.
Since the default with emails is generally filing in personal folders (which
are hard to find and access by other than the person concerned), copying
them to the job file is specified (Procedure 2.4). Incidentally, the ‘job file’
is of course a network folder and hard copies of records or physical files
are generally not needed.
procedure may reflect practice that was well established by ARSP before
implementing BS ISO 20252 and covers the relevant requirements for
telephone interviewers as per Clause 5 of the standard. However, this is
an area where some companies may need to make significant changes to
established practices and quite possibly changes with major cost
implications.
ARSP procedures up to this point (Procedure 10) cover processes that are
likely to be familiar to all research companies regardless of whether or
not there is a formal quality system. Procedure 11, however, only makes
sense as part of a quality system and meets requirements of Clause 3 of
BS ISO 20252 that are unlikely to reflect established practice. At its heart
is the need for making process improvements and dealing with identified
problems. If Procedure 11.2 is followed with commitment, rather than
just going through the motions to meet the standard, this alone can
justify the cost and effort of implementing a formal quality system. The
philosophy reflected is that of honestly facing that mistakes and
problems happen and then seeking to understand their causes and
devising solutions to lessen the chances of reoccurrence; the aim is
preventive as well as corrective action. If a problem occurs in a project,
for example one interviewer’s work is found to be defective at the end
of the fieldwork period, something needs to be done about it there and
then (perhaps allocate the quota to another interviewer and possibly
extend the fieldwork period) and this is appropriate corrective action.
However, if nothing else is done, there is a good chance the same
problem will happen on another project. A longer term, preventive
action (e.g. re-training the interviewer) is needed as well. Procedure 11.2
provides mechanisms for addressing problems and process improvement.
As in other procedures, responsibilities are assigned and records are kept.
Again, though, we must emphasize that Procedure 11.2 is just one way of
meeting the requirements of the standard. Other ways of working could
be equally effective, meet the requirements and better fit how another
company works.
The ‘find problems and fix them’ philosophy also includes deliberately
seeking them out. Internal auditing of the quality system is a tool to do
just this and is a specific requirement of BS ISO 20252 (Subclause 3.6.3).
Procedure 11.1 is to ensure effective internal audits are carried out. More
Example of appendices
The final parts of the specimen procedures are two appendices. Appendix
1 is an adjunct to Procedures 1 and 7 and respectively provides checklists
of the content of proposal and reporting coverage. In each case, the
items listed are as per the BS ISO 20252 requirements in this area that are
relevant to the types of project carried out by ARSP. These lists are
needed but equally could be included as part of the two procedures and
having them here is just a matter of convenience; proposal and reporting
drafters may use the appendix as a stand-alone prompt card. Appendix 2
is not strictly a requirement of the standard. However, it serves two
purposes: to guide outside assessors and to check that the system meets
BS ISO 20252. If certification is sought (see below) the assessors will find
it useful to have a brief summary of the company and its implementation
of the standard.
61
An organization chart is a requirement of BS ISO 20252:2012 and the document should be
controlled in the same way as procedures. Given that organizations often change it may be
convenient to have the chart separately rather than having to revise a fuller procedure
document.
probably be most involved in this system launch but senior and line
managers need to be seen to be committed as well. It is not unusual for
senior staff to find personal difficulties in following a system but ‘do as I
say not what I do’ is no good.
Auditing the whole system and all processes could be done together but
it is better to do a bit at a time, particularly when auditors and the
quality manager have to fit the time required around their ‘day jobs’.
The quality manager, therefore, needs a plan of when each element will
be audited and this should be documented in some form (outside
assessors will want to see the plan). While annual internal audits of the
whole system may be adequate once a quality system is well embedded,
this will not be sufficient to start with.
First audits need to be carried out soon after the launch date and
repeated until the system appears to be working effectively; the results
of initial audits will determine this. During early audits project records
may be limited because the projects have not gone through all or most
processes since the quality system start date, but this is not a reason to
delay auditing for too long. At the individual audit level (i.e. a session)
the auditors also need to plan what in detail they are going to cover,
where and who they will need to see. Some sort of checklist for each
audit, based on the relevant procedures, is needed. Appointments may
need to be made with the ‘auditees’ to ensure availability; unexpected
visits may not be practical (staff may be away from the office) or
desirable.
Although the auditing role can be wider, the main objective should be
comparing what the procedures say should happen with evidence of
actual practice. For example specimen Procedure 9 states:
Is there evidence that this happens in the phone unit? For the most
recently recruited interviewers, can the documentary evidence be
produced? Checking this will involve a meeting with the phone unit
manager or a supervisor and requesting retrieval of these records. To
some extent the evidence may be verbal but good procedures must be
auditable via the review of records of some sort and not dependent on
hearsay alone. In the example, it may be the case that some new
interviewers were monitored but others not, so there is a need for at
least a sample to be taken. This is even more the case when the evidence
sought is in project records. Practice is required to be consistent across
projects so establishing whether, for example, questionnaires are
‘signed-off’ requires looking at a number of job files and the sample size
Audits are not meant to blame anyone but if a reason is given for
non-compliance this can usefully be noted as well; such comments may
well suggest the solution, e.g. ‘I did not know I was meant to monitor
first job’. Observations are other notes the auditor may think it useful to
make – perhaps problems less than non-conformities, e.g. it took a long
time to find the monitoring records. The audit report can then be
discussed with the quality manager who should have the responsibility, in
consultation with other staff including those directly affected, of finding
solutions to any problems. The emphasis needs to be longer term; finding
a solution so that the problem is less likely to occur again. Incidentally, if
the non-compliance found is a failure to complete a required record,
making the record retrospectively serves no purpose at all.
Revising procedures
The two main ways of solving non-compliance with procedures is training
and procedure revision. Particularly when the quality system is first
implemented, the most common reason for staff not following
procedures is simply that they did not know what they were supposed to
do and this will be found regardless of whatever effort is made in
launching and communicating the system. Often the fact that the
non-compliance has been identified, through an audit, effectively solves
the problem; the staff members concerned now know what they should
be doing. Sometimes, though, a more formal training session will be
needed and as a last resort the line manager will need to be pushed to
ensure future conformity.
Less commonly, the reason for non-compliance may be that the particular
procedure does not work well or is ambiguous. In this case the solution is
In the first few months after implementing the quality system, the need
to revise some procedures should be expected, a matter of learning
through experience. Changes in the documented quality system, however,
need to be made in a formal way to ensure that at any one time there is
only one current version to be followed. The specimen procedures
include (in 11.3) a procedure for revision, including authority to make the
change (the procedure could just as well allow the quality manager to
make the revision without reference to anyone else) and a record of
what was changed (a copy of the superseded procedure). All staff
members need to be made aware that procedures have been revised; do
not just hope they will find this out otherwise.
An organization may also decide to revise the quality system apart from
dealing with problems found in internal audits. BS ISO 20252 requires
problem solving and a quality improvement process far wider than
reacting to internal audits. There is a need to consider process-related
problems (e.g. interviewer error rates), issues raised by clients (complaints
or as feedback from monitoring client satisfaction) or deficiencies in the
work of subcontractors. All positive suggestions for improvement should
also be considered. A process is needed for this as well, such as 11.2 of
the specimen procedures.
62
To clear up some confusion of terminology, it is assessment bodies which are accredited
and they assess companies by auditing to standards such as BS ISO 20252:2012 and then
certify or register the companies. Although it is a commonly-used phrase, companies are
not accredited; they are certified.
63
The general standards for assessment are BS ISO/IEC 17021:2011, and ISO/IEC Guide
28:2004.
Assessment bodies have a positive attitude. They are looking for evidence
of compliance so that certification can be awarded. They are not there to
find problems. However, in a full assessment audit, it is likely that
instances of non-compliance with the company’s own system and
BS ISO 20252 will be found. No one and no company is perfect.
Non-compliances are classed, by assessors, as either ‘major’ or ‘minor’
non-conformities. A major non-conformity is where a significant part of
the quality system and standard is not met (e.g. interviewer validation is
not done to the required level) and where this is the case the assessment
body will be unable to certify the company until the problem is solved
and this will normally require an additional audit visit (with extra
charges). Really, no company should be ‘failed’ in this way; adequate
internal auditing should have identified the problem and action should
have been taken before outside assessment. Minor non-conformities are,
for example, specific omissions in following procedures in isolated cases,
and in an assessment a number of such may well be found. Unless such
minor non-conformities are so numerous as to be judged collectively as a
major non-conformity, the assessors will still recommend certification
providing the company agrees to solve the problems before the next
routine assessment visit. This normally involves the formality of having a
‘corrective action request’ document raised by the assessor and at the
next visit, resolution of this will be the first thing checked. Give or take
any minor non-conformities and corrective action requests, the assessor
will at the end of the audit indicate whether or not certification will be
recommended; in the positive case the actual certificate will be issued
shortly by a senior officer of the assessment body.
64
BS ISO 20252:2012 includes many requirements which are specific to market research
processes, whereas BS EN ISO 9001 is a far more general standard which is applicable to
almost any organization regardless of their business.
Specimen procedures
ARSP Ltd. quality system procedures
Introduction
The quality system meets the requirements of: BS ISO 20252 – see
Appendix 2.
All staff are required to follow the procedures relevant to the work
they carry out.
Specimen procedures
Specimen procedures
2. Commission of projects
Specimen procedures
• All project details (including that the work is being carried out,
the methodology and results) are confidential to the client and
are not given to third parties without the written agreement of
the client.
• The identity of the client is not revealed to research
respondents unless this is agreed in writing with the client
and/or there is a legal obligation to do so.
• All project records are kept securely and available to ARSP staff
only as required to carry out work.
3. Project planning
Specimen procedures
4. Qualitative research
Specimen procedures
5. Data collection
Specimen procedures
Specimen procedures
6. Data processing
Specimen procedures
7.1 Reporting
Specimen procedures
7.5 IT security
• All files on the network servers are backed up daily with copies
retained off-site.
• Full and continually updated virus protection and firewall
programmes are in place – all files in the network are scanned
for viruses.
• The computer network, including hardware and software, is
supported and maintained and these arrangements are
documented.
8. Sub-contracting
Specimen procedures
• The research team will, on completion, review the work for the
project of the sub-contractor.
• The team leader decides whether to request access to the
sub-contractor’s project records to confirm compliance with
BS ISO 20252.
• Any significant problems with the sub-contractor’s work are
reported in writing to the quality manager. The project director
considers the need for any immediate action.
Specimen procedures
Specimen procedures
Specimen procedures
• The quality manager ensures that the copies of the MRS Code
of Conduct and associated guidelines, accessible on the
network, are up to date.
Specimen procedures
ARSP carries out full service research projects for clients. Most
projects are quantitative in scope with data collection by telephone
interviews. Qualitative research is also carried out but with
respondent recruitment sub-contracted. Quantitative data
collection, other than by telephone interviewing, is also
sub-contracted.
Specimen procedures
Specimen procedures
Also listed are other sources consulted but which are not referenced.
British Quality Foundation (2010), The EFQM Excellence Model 2010, BQF,
London
EFAMRO (2010) Guidance Note: Data Protection in the EU, EFAMRO, The
Hague
Institute for Employment Studies (2008), ‘Does IIP add value to business?’,
Employment Studies newsletter No 7, IES, Brighton, UK
Investors in People (2011) It’s Your Choice: the Standards and Beyond –
The Investors in People Framework, The Stationary Office, London
Juran, J M (1989) Juran on Leadership for Quality, Free Press, New York
May, J (2006) ‘Ladders, Stars and Triangles: Old and New Theory for the
Practice of Public Participation’, International Journal of Market Research,
48(3)
Moore, W E (1970) The Professions: Roles and Rules, Russell Sage, New
York
MRS (2003) The Data Protection Act 1998 & Market Research: Guidance
for MRS Members, MRS, London
MRS (2011) The Annual Survey of the UK Market Research Industry 2010,
MRS, London
O’Brien, J (1987) ‘Two answers are better than one’, Journal of the
Market Research Society, 29(3)
Poynter, R and Lawrence, G (2007) ‘Insight 2.0: new media, new rules,
new insight’, Conference proceedings, ESOMAR
If you found this book useful, you may also want to buy:
Complaints should be an easy way for your customers to express an opinion about
your products and services. This book has been designed to help any company or
organization that wants to improve the effectiveness of customer complaint
processes and systems and also adhere to the principles of the complaints
management standard, ISO 10002:2004. The book draws on case studies from a
variety of sectors, such as financial services, healthcare, not-for-profit, utilities, and
travel, to present the reader with opportunities to reflect on good practice, while
realizing the underlying benefits of the standard. There are quotes and statistics that
can support an internal business case, as well as practical checklists for handling
complaints, analysing root causes, and conducting system reviews. The book also
references the benefits that can be delivered from external dispute resolution
schemes and how ISO 10003:2007 delivers a standard for improving the
effectiveness of such schemes.
The book is available in two formats – text alone and text with standards.
A5 paperback · ISBN 978 0 580 71876 2/ (with standards) 978 0 580 76553 7
· 247/344pp · £40.00 or £160 with standards (member)/ £288 (non-member)
· BSI order reference: BIP 2211/2222
For more details see http://shop.bsigroup.com/bip2211
Forthcoming
Raising the Standard for Collaboration
David Hawkins
The aim of the book is to provide an introduction to the concepts and value of
collaboration and the standard BS 11000-1:2010, Collaborative business
relationships. A framework specification. It focuses on providing a first step to
adoption and implementation and is a low cost introduction before venturing into the
purchase of the standard and implementation investment.