TEM-515: STATISTICAL

ANALYSES WITH COMPUTER

APPLICATION

Lecture Series 1: Introduction to Data Collection

Lecture 2: Variables and Levels of Measurement

Instructor: Dr. Zahara Batool Date: 04 February 2015

Lecture 2 Content Outline
 Characteristics of statistics

 Variables
 Types of variables
 Level of Measurement
 Errors of Measurement
 Univariate, bivariate, and multivariate data sets
Variables
 The numbers which we observe and record in a data matrix
may have very different meanings and this fundamentally
affects the inferences we can draw from them.
 A variable is a characteristics or property of a person, an
object or a situation, comprising a set of different values or
categories.
 For example, height, weight, age, blood type, gender,
nationality
Types of Variables
 Blood type
Record in terms of kind –
 Gender
Qualitative Variables
 Nationality

 Height,
 Weight Measured in units –
Quantitative Variables
 Age
A. Qualitative variable – Categorical
Variables
 Categorical variables (CVs) arise when individuals are
allocated to categories. Country of birth is a CV as is the
highest level of education.
 Numbers can be used to code the categories into which
individual falls but those numbers are nothing more than
codes.
Qualitative variable – Categorical
Variables
However, if for example, respondents are asked to say
whether they strongly agree’’, ‘agree’, or ‘disagree’ or ‘strongly
disagree’ with some proposition, it is not unusual for equally
spaced scores such as 1, 2, 3, 4 to be assigned to these
categories, thus appearing to turn them into metrical variables.
 Ordered Categorical Variables e.g. level of education
 Unordered Categorical Variables e.g. gender
B. Quantitative Variable – Metrical
Variable
Metrical variables are those which can be recorded on some
kind of scale, like response times, lengths or examination
sores, where the interval b/w two values has a quantitative
interpretation.
 Continuous Variable can take any value in an interval e.g. ?

 Discrete Variable has only particular values e.g. ?

 Discrete and Continuous Variables

Discrete

Countable or Measureable– Qualitative Variables

Continuous
Discrete Variable
 Variable that can take only a discrete set of integers or whole
numbers i.e., the values are taken by jumps or break.
 It represents a count data.
 Example:

the number of persons in a family, the number of rooms in a

house, the number of deaths in an accident, the income of an
individual
Continuous Variable
 It is a variable that can take on any value – fractional or
integral – within a given interval, i.e. its domain is an interval
with all possible values without gaps.
 It represents measurement data.
 Example:

the age of a person, the height of a plant, the weight of a

commodity, the temperature at a place
 Observable and Latent Variables
Variables can also be classified under an entirely different
classification scheme in order to perform analyses such as factor
analysis and related methods.
This classification is concerned with whether or not the variable can
be observed.
 Manifest or observable variables can be observed directly.
 However, latent or hidden variables cannot (underlying
variable).
 Sometimes, a variable can only be observed partially, and
described as incompletely observed.
Levels of Measurement –
Measurement Scale
A data set is a collection of numerical observations or
variables – whether qualitative or quantitative. Therefore, it is
useful to identify their level of measurement.
 Nominal (the most simple)

 Ordinal

 Interval

 Ratio (the most sophisticated)

Nominal – Unordered Categorical Variable
 Nominal is hardly measurement. It refers to quality more than quantity.

 Thus, a nominal variable is the same as an unordered categorical

variable.
 For example, a nominal level of measurement is simply a matter of
distinguishing by name, e.g., 1 = male, 2 = female. Even though, the
numbers 1 and 2 are used, they do not denote quantity.
 The binary category of 0 and 1 used for computers is a nominal level
of measurement - categorical levels of variables.
Nominal Data Examples
 MEAL PREFERENCE:

Breakfast, Lunch, Dinner

 POLITICAL ORIENTATION:

Republican, Democratic, Libertarian, Green

 RELIGIOUS PREFERENCE:

1 = Buddhist, 2 = Muslim, 3 = Christian, 4 = Jewish, 5 =

Other

Nominal time of day 

Categories
(no additional information)
AM PM
Ordinal – Ordered Categorical Variable
 Ordinal refers to order in measurement.
 An ordinal scale indicates direction, in addition to providing nominal
information.
 For example, Low/Medium/High or Faster/Slower are examples of ordinal
levels of measurement.
 Ranking an experience as a "nine" on a scale of 1 to 10 tells us that it
was higher than an experience ranked as a "six." Many psychological
scales or inventories are at the ordinal level of measurement.
 Ranks are assigned as the lowest level measurement than scale data.
Ordinal Data Examples
 RANK:

1st place, 2nd place, ... last place

 LEVEL OF AGREEMENT:

No, Maybe, Yes

 POLITICAL ORIENTATION:

Left, Center, Right

Ordinal time of day

Indicates direction or order of occurrence
(spacing between is uneven)
Night Dawn Noon Afternoon Evening
Scale Data – Metrical Variables
 At the highest level, scale data are measurements of quantity in
units on an independent scale.
 Heights and weights are obvious examples. So also are
performance scores, such as the number of times a participant
hits a target, as well as IQs, responses to questionnaires and
other psychometric data.
 In such a data set, each individual score or datum carries
information independently of the other data.
 Interval scale Data
 Ratio scale Data
Interval Scale Data
 Interval data is like ordinal data except we can say the
intervals between each value are equally split.
 The most common example is temperature in degrees
Fahrenheit. The difference between 29 and 30 degrees is the
same magnitude as the difference between 78 and 79.
 With attitudinal scales and the Likert questions you usually
see on a survey, these are rarely interval, although many
points on the scale likely are of equal intervals.
Interval Scale Data Example
 TIME OF DAY on a 12-hour clock

Interval time of day 

Equal intervals; Analog (12 –hr)
(clock , difference between 1 and 2 pm is same as difference between
11 and 12 am)
12 1 2 3 4 5 6 7 8 9 10 11 12
Ratio Scale Data
 In addition to possessing the qualities of nominal, ordinal, and interval
scales, a ratio scale has an absolute zero. Ratio level variables are
necessarily non-negative.
 Amounts of money, length and weight are familiar examples.

 The units in which they are measured are arbitrary - money may be
measured in dinars or dollars, for example.
 They are termed ratio level variables because ratios are independent of
the units of measurement.
 For example, a camera which costs 3 times as much as another when
priced in dollars will still be 3 times as costly if the ratio is expressed in
any other currency.
 Ratio Scale Data Example
 RULER: inches or centimeters
 YEARS of work experience

 INCOME: money earned last year

 NUMBER of children

Time of day on Ratio Scale

24 – hour time has an absolute 0 (midnight);
(14 o'clock is twice as long from midnight as 7 o'clock)
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 1 1 1 1 1 20 21 22 23 24
5 6 7 8 9
Remember:
 Measurement at the interval or ratio level is desirable because the more
powerful statistical procedures available for Means and Standard Deviations.
 To have this advantage, often ordinal data are treated as though they were
interval; for example, subjective ratings scales (1 = terrible, 2= poor, 3 = fair,
4 = good, 5 = excellent).
 The scale probably does not meet the requirement of equal intervals – it is
not known whether the difference between 2 (poor) and 3 (fair) is the same
as the difference between 4 (good) and 5 (excellent).
 In order to take advantage of more powerful statistical techniques,
researchers often assume that the intervals are equal.
Univariate, bivariate and
multivariate data sets
 In a univariate data set, all the data refer to just one variable.
 In a bivariate data set, there are measurements on two
variables.
 In a multivariate data set, there are measurements on three
or more variables.
Univariate Data Set
The three most important properties of a univariate data set are;
 The average (as measured by the mean, the median, the mode or
some other measure of central tendency)
 The spread or dispersion of the scores about the average, as
measured by the standard deviation, the variance or a range
measure.
 The shape of distribution (symmetric, normal, skewed and so on).
Bivariate Data Set
 With a bivariate distribution, interest shift to the possibility of
a statistical association, or correlation, b/w the variables.
 Do people with higher incomes tend to score higher on
psychometric tests? Is a number of years of formal education
an important factor in person’s income?
Multivariate Data Set
 With a multivariate data set, the researcher is often
interested in the extent to which certain variables can be
accounted for or explained in terms of others in the data set.
 How well person’s income can be predicted from number of
years of formal education, psychometric intelligence and
parents’ income. These are problems in regression.
Errors of Measurement
Continuous variables can never be measured with perfect fineness because …

Absolute error

Relative error Departure from the true value –

Error of Measurement
Percentage error

Error has both magnitude and direction and error in statistics does not mean
mistake which is a chance inaccuracy
Errors of Measurement

Biased error

Unbiased error Types of Error

Accurate measurement
 Biased Error

when observed value is consistently and constantly higher

or lower than the true value – cumulative or systematic errors
 Unbiased Error

When the deviations, i.e. the excess and defects, from the
true value tend to occur equally often – random error or
accidental errors
 Accurate Measurement

A measurement free from all classes of errors.

Use and Misuse of Statistics
‘‘There are three kind of lies – lies, damned lies, and
statistics.’’

Benjamin Disraeli,
Former British Prime Minister
A Word of Caution
Lecture 2: Variables and Levels of
Measurement

Thank you for listening