Professional Documents
Culture Documents
Types of questions:
2. Unambiguous
It should be worded so that the audience interprets the answer the way they are
meant to be interpreted and not some other way cuz that will decrease the accuracy
of the result
LAYOUT:
Introductions
- at beginning, at the beginning of sections
-explain the purpose
-state the organisation behind it
-confidentiality/anonymity
-data security
-contact details
Instructions
should be clear, unambiguous
should be distinct from the questions
E.g. How often do you buy any type of fish for your family? Please
circle the appropriate number.
Appearance
Layout
Colour
Typeface
Length
FLOW OF QUESTIONS
AVOID;
• Ordering items in such a way that necessitates referral back to earlier questions
• Conditioning of later responses by earlier ones
• Preceding knowledge questions with other questions giving the answers
VI. Evaluating the Questionnaire and Designing for Analysis
Approaching respondents
Advance warning
• Explanation of selection
• Sponsorship
• Envelope
• Incentives
• Publicity
• Confidentiality
• Anonymity
• Follow-ups
Experiments
Experiment’s validity
So how are we going to know that our experiment is a valid one and the relationship
that we claim exists? Two considerations will affect an experiment’s validity,
2. The influence of extraneous factors – extraneous factors are factors other than
the independent and dependent variables which may influence the results of the
experiment.
1. Pre-experimental designs – These are designs that don’t control for extraneous
factors by randomisation.
4. Statistical designs – These are designs that allow for the statistical control and
analysis of external variables.
PRE-EXPERIMENTAL DESIGN
One shot case study – This is where a group of respondents are exposed to a
treatment and a single measure of the dependent variable is taken. E.g. taking you
exposing you to a film on fur trade and asking you about your attitudes after. You
can’t draw valid conclusions from this for a number of reasons,
i. We don’t know what your attitudes to the fur trade were before the film,
ii. The sample is not randomly selected, the researcher has selected the group and
younger people might be more affected than older age groups. This design lacks
control for extraneous variables, history, maturation, selection, mortality and
regression.
One Group pre-test-post-test design – In this design the group of respondents are
measured twice, so going back to the fur film the researcher would measure your
attitudes before showing you the film and after showing you the film. This allows the
researcher to calculate the effect of the treatment in this case the film. Once again
this does not really take any extraneous variables into account.
Static Group Design – This is a two group experimental design. One group is called
the experimental group and one is called the control group. Measurement is made
on the groups after the experimental group has been given their treatment. So you
would be split into two groups, one would watch the film and then the attitudes of
both groups would be measured. Problems because the respondents are not
randomly assigned to groups so they may differ before the treatment.
2. Post-Test-Only Control Group Design – The post-test only design does not
involve any pre-measurement. It’s assumed that the respondents in each group are
similar as they have been randomly assigned to each group.
3. Solomon Four Group Design – This design not surprisingly involves four
groups, there is the experimental group and control group both are pre-
measured, control groups 2 and 3 are not pre-measured. The experimental group
has the treatment as does control group 2 then all four groups are measured.
Control group 2 controls for sensitising which pre-testing may do. Control group 3
controls for history, any change which may occur over time, between measures.
Quasi-Experimental Designs
Time series design – this involves taking a series of measurements for a group of
respondents. The treatment might be administered by the researcher but may occur
naturally. After the treatment periodic measurement continues to take place.
Maturation is partially controlled as although it may occur over time it may not
effect O5 and O6. If the respondents are selected randomly, selection bias can be
reduced. The major weakness is that as it is occurring over time it does not control
for history. Mortality may be a problem so you might need to offer incentives to the
participants.
Multiple Time Series Design – The multiple time series is similar to the time
series, but another group of respondents are used to serve as a control group.
Limitations to Experimentation
• Time
• Cost
• Administration
SURVEYS
The survey method involves obtaining information using structured questions
designed to elicit specific information from respondents. The questions may cover
the respondent’s behaviour, intentions, attitudes, awareness, motivations and
demographic and lifestyle characteristics. The data is structured in that the questions
are prepared and gone through with each of the respondents.