Introduction to SPSS

Session 2
Creating a data file and entering data

Step 1. Enter the data—that is, the values obtained from each
participant or respondent for each variable.

Step 2. The next step is to set up the structure of the data file by
‘defining’ the variables.
Screening and Cleaning the data

Step 1: Checking for errors. First, you need to check each of your
variables for scores that are out of range (i.e. not within the range of
possible scores).

Step 2: Finding and correcting the error in the data file. Second, you
need to find where in the data file this error occurred (i.e. which case is
involved) and correct or delete the value.
Procedure for checking categorical variables
From the main menu at the top of the screen, click on Analyze, then click on Descriptive Statistics,
then Frequencies.

Choose the variables that you wish to check (e.g. sex, marital, educ).

To assist in finding the variables you want you can right click on the list of variables and select Sort
Alphabetically, or to show Variable Names or Variable Labels.

Click on the arrow button to move these into the Variable box.

Click on the Statistics button. Tick Minimum and Maximum in the Dispersion section.

Click on Continue and then on OK.

Checking continuous variables

From the menu at the top of the screen, click on Analyze, then click on Descriptive
statistics, then Descriptives.

Click on the variables that you wish to check. Click on the arrow button to move them
into the Variables box (e.g. age).

Click on the Options button. You can ask for a range of statistics. The main ones at this
stage are mean, standard deviation, minimum and maximum. Click on the statistics you
wish to generate.

Click on Continue, and then on OK

Finding And Correcting The Error In The
Data File
1. Click on the Data menu and choose Sort Cases.

2. In the dialogue box that pops up, click on the variable that you
know has an error (e.g. sex) and then on the arrow to move it into
the Sort By box. Click on either ascending or descending (depending
on whether you want the higher values at the top or the bottom).

3. Click on OK.
Preliminary analyses

In studies involving human participants, it is useful to collect information on

the number of people or cases in the sample, the number and percentage of
males and females in the sample, the range and mean of ages, education
level, and any other relevant background information.

Prior to doing many of the statistical analyses (e.g. t-test, ANOVA,

correlation), it is important to check that you are not violating any of the
‘assumptions’ made by the individual tests.
CATEGORICAL VARIABLES

Procedure for obtaining descriptive statistics for categorical variables

1. From the menu click on Analyze, then click on Descriptive Statistics, then Frequencies.

2. Choose and highlight the categorical variables you are interested in (e.g. sex). Move
these into the Variables box.

3. Click on OK

Note: If you have very unequal group sizes, particularly if the group sizes are small, it
may be inappropriate to run some of the parametric analyses (e.g. ANOVA).
CONTINUOUS VARIABLES

1. From the menu click on Analyze, then select Descriptive Statistics, then Descriptives.

2. Click on all the continuous variables that you wish to obtain descriptive statistics for.
Click on the arrow button to move them into the Variables box (e.g. age, Total perceived
stress: tpstress).

3. Click on the Options button. Make sure mean, standard deviation, minimum,
maximum are ticked and then click on skewness, kurtosis.

4. Click on Continue, and then OK

Positive skewness values suggest that scores are clustered to the left at
the low values.

Negative skewness values indicate a clustering of scores at the high end

(right-hand side of a graph).

Positive kurtosis values indicate that the distribution is rather peaked

(clustered in the centre), with long thin tails.

Kurtosis values below 0 indicate a distribution that is relatively flat (too

many cases in the extremes).
 ASSESSING NORMALITY
1. From the menu at the top of the screen click on Analyze, then select Descriptive Statistics, then Explore.

2. Click on the variable(s) you are interested in (e.g. Total perceived stress: tpstress). Click on the arrow button to move

them into the Dependent List box.

3. In the Label Cases by: box, put your ID variable.

4. In the Display section, make sure that Both is selected.

5. Click on the Statistics button and click on Descriptives and Outliers. Click on Continue.

6. Click on the Plots button. Under Descriptive, click on Histogram to select it. Click on Stem-and-leaf to unselect it. Click

on Normality plots with tests. Click on Continue.

7. Click on the Options button. In the Missing Values section, click on Exclude cases pairwise. Click on Continue and then

OK
In the table labelled Tests of Normality, you are given the results of the
Kolmogorov-Smirnov statistic. This assesses the normality of the
distribution of scores. A non-significant result (Sig. value of more
than .05) indicates normality.
CHECKING FOR OUTLIERS

First, have a look at the Histogram. Look at the tails of the distribution.
Are there data points sitting on their own, out on the extremes?

Second, inspect the Boxplot. Any scores that IBM SPSS considers are
outliers appear as little circles and asterix with a number attached (this
is the ID number of the case).
Using graphs to describe and explore the data

HISTOGRAMS

distribution of a single continuous variable (e.g. age, perceived stress

scores).
1. From the menu click on Graphs, then select Chart Builder. Click OK.

2. To choose the type of graph that you want, click on the Gallery tab, and choose Histogram.

3. Click on the first image shown (Simple Histogram) and drag it up to the Chart Preview area, holding your left mouse button

down.

4. Choose your continuous variable from the list of Variables (tpstress) and drag it across to the area on the Chart preview screen

labelled X-Axis holding your left mouse button down. This will only work if you have identified your variable as Scale in the

Data Editor window (the icon next to the variable should be a ruler).

5. If you would like to generate separate graphs for different groups (e.g. male/female) you can click on the Groups/Point ID tab

and choose Column Panels variable option. This will produce separate graphs next to each other; if you would prefer them to

be on top of one another choose the Rows panel variable.

6. Choose your categorical grouping variable (e.g. sex) and drag it across to the section labelled Panel in the Chart Preview area.

7. Click on OK
BAR GRAPHS

The bar graph can show the number of cases in particular categories, or
it can show the score on some continuous variable for different
categories.
1. From the menu at the top of the screen, click on Graphs, then select Chart Builder and click

2. OK. Click on the Gallery tab and click on the second graph displayed (Clustered Bar). Holding your
left mouse button down drag this graph to the Chart Preview area.

3. From the Element Properties window click on Display error bars, and then on the Apply button at
the bottom of the window. If for some reason you don’t have the Element Properties window click
on the Element Properties box on the far right of the main screen.

4. From the list of Variables drag one of your grouping variables (e.g. sex) to the section on the Chart
Preview screen labelled Cluster on X: set colour. Click and drag your other categorical variable (e.g.
agegp3) to the section labelled X-Axis at the bottom of the graph. Click and drag your continuous
variable (Total Perceived Stress: tpstress) to the remaining blue section, the Y-axis.

5. Click on OK
LINE GRAPHS

A line graph allows you to inspect the mean scores of a continuous

variable across a number of different values of a categorical variable
(e.g. time 1, time 2, time 3).
1. From the menu at the top of the screen, select Graphs, then Chart Builder, and then
OK.
2. Click on the Gallery tab and then click on the second graph shown (Multiple Line). Drag
this option to the Chart preview area holding your left mouse button down.
3. From the Variables list drag your continuous variable (Total perceived stress:tpstress)
to the Y-axis. Drag one of your categorical variables (e.g. sex) to the section labelled Set
color and drag the other categorical variable (agegp5) to the X-Axis. If you want to display
error bars you can request this from the Element Properties window—tick the box
Display error bars and click the Apply button at the bottom of the screen.

4. Click on OK
SCATTERPLOTS

Scatterplots are typically used to explore the relationship between two

continuous variables (e.g. age and self-esteem).
1. From the menu at the top of the screen, click on Graphs, then Chart Builder, and then OK.

2. Click on the Gallery tab and select Scatter/Dot. Click on the second graph (Grouped
Scatter) and drag this to the Chart Preview area by holding your left mouse button down.

3. Click and drag your continuous independent variable (Total PCOISS:tpcoiss) to the X-Axis,
and click and drag your dependent variable (Total perceived stress:tpstress) to the Y-Axis.
Both of these variables need to be nominated as Scale variables. If you want to show groups
(e.g. males, females) separately choose your categorical grouping variable (e.g. sex) and drag
to the Set Colour box.

4. Click on OK
BOXPLOTS

Boxplots are useful when you wish to compare the distribution of

scores on variables.
1. From the menu at the top of the screen, click on Graphs, then select Chart
Builder, and click OK.

2. Click on the Gallery tab and choose the first graph displayed (Simple boxplot).
Drag it up to the Chart Preview area, holding your left mouse button down.

3. From the Variables box choose your categorical variable (e.g. sex) and drag it to
the X-Axis box on the Chart Preview area. Drag your continuous variable (Total
Positive Affect:tposaff) to the Y-axis.

4. Click on the Groups/Point ID and select Point ID label.

5. Click on OK
Manipulating the data

• Adding up the scores from the items

• Collapsing continuous variables (e.g. Age) into categorical variables

• Reducing or collapsing the number of categories of a categorical variable

• Recoding data entered as text (male, female) to numeric data (1, 2)

• Transforming skewed variables for analyses that require normally

distributed scores
CALCULATING TOTAL SCALE SCORES

Step 1: reverse any negatively worded items.

Step 2: add together scores from all the items that make up the
subscale or
Create new variables
1. From the menu at the top of the screen, click on Transform, then click on Recode Into Different Variables.

2. Select the items you want to reverse (op2, op4, op6). Move these into the Input Variable— Output Variable

box.

3. Click on the first variable (op2) and type a new name in the Output Variable section on the right-hand side of

the screen and then click the Change button.

4. Click on the Old and New Values button.

In the Old Value section, type 1 in the Value box.

In the New Value section, type 5 in the Value box (this will change all scores that were originally scored as 1 to a

5).
5. Click on Add. This will place the instruction (1 → 5) in the box labelled Old > New.

6. Repeat the same procedure for the remaining scores. For example: Old Value—type in

2 New Value—type in 4 Add

Old Value—type in 3 New Value—type in 3 Add

Old Value—type in 4 New Value—type in 2 Add

Old Value—type in 5 New Value—type in 1 Add

Always double-check the item numbers that you specify for recoding and the old and

new values that you enter.

7. Click on Continue and then OK

Adding up the total scores for the scale
1. From the menu at the top of the screen, click on Transform, then click on Compute Variable.

2. In the Target Variable box, type in the new name you wish to give to the total scale scores.

3. Click on the Type and Label button. Click in the Label box and type in a description of the scale
(e.g. total optimism). Click on Continue.
4. From the list of variables on the left-hand side, click on the first item in the scale (op1). Click on
the arrow button to move it into the Numeric Expression box.

5. Click on + on the calculator.

6. Click OK
Collapsing continuous variables (e.g. Age)
into categorical variables
1. From the menu at the top of the screen, click on Transform and choose Visual Binning.
2. Select the continuous variable that you want to use (e.g. age). Transfer it into the Variables to Bin box. Click
on the Continue button.
3. In the Visual Binning screen, a histogram showing the distribution of age scores should appear.
4. In the section at the top labelled Binned Variable, type the name for the new categorical variable that you will
create (e.g. Agegp3).
5. Click on the button labelled Make Cutpoints. In the dialogue box that appears, click on the option Equal
Percentiles Based on Scanned Cases. In the box Number of Cutpoints, specify a number one less than the
number of groups that you want (e.g. if you want three groups, type in 2 for cutpoints). In the Width (%)
section below, you will then see 33.33 appear. This means that IBM SPSS will try to put 33.3 per cent of the
sample in each group. Click on the Apply button.
6. Click on the Make Labels button back in the main dialogue box. This will automatically generate value labels
for each of the new groups created.
7. Click on OK
Reducing or collapsing the number of
categories of a categorical variable
1. From the menu at the top of the screen, click on Transform, then on
Recode into Different Variables.

2. Select the variable you wish to recode (e.g. educ). In the Name box, type a
name for the new variable that will be created (e.g. educrec). Type in an
extended label if you wish in the Label section. Click on the button labelled
Change.

3. Click on the button labelled Old and New Values.

4. In the section Old Value, you will see a box labelled Value. Type in the first code or value of your
current variable (e.g. 1). In the New Value section, type in the new value that will be used (or, if the same
one is to be used, type that in). In this case I will recode to the same value, so I will type 1 in both the Old
Value and New Value sections. Click on the Add button.

5. For the second value, I would type 2 in the Old Value but in the New Value I would type 1. This will
recode all the values of both 1 and 2 from the original coding into one group in the new variable to be
created with a value of 1.

6. For the third value of the original variable, I would type 3 in the Old Value and 2 in the New Value. This
is just to keep the values in the new variable in sequence. Click on Add. Repeat for all the remaining
values of the original values. In the table Old > New, you should see the following codes for this example:
1→1; 2→1; 3→2; 4→3; 5→4; 6→5.

7. Click on Continue and then on OK

USING AUTOMATIC RECODE TO
CONVERT TEXT TO NUMERIC VALUES
1. Click on Transform and then select Automatic Recode.

2. Select your text variable (e.g. Sex) and move this into the Variable-New Name
box.

3. Type the name you would like to give the converted variable in the New name
box (e.g. SexNum). Click on the Add New Name button.

4. Click on OK.
Reliability and Validity
Reliability Validity
Definition The consistency and stability of The accuracy and relevance of a
a measurement. measurement.
Purpose To ensure that a measurement To ensure that a measurement
produces consistent results accurately measures what it is
over time and across different intended to measure.
situations.
Does all the items of the scale Does the scale have a
have high correlation. theoretical underpinning.
Checking the reliability of a scale
One of the most commonly used indicators of internal consistency is Cronbach’s alpha coefficient. Ideally, the Cronbach
alpha coefficient of a scale should be above .7 (DeVellis 2012).

1. From the menu at the top of the screen, click on Analyze, select Scale, then Reliability Analysis.

2. Click on all of the individual items that make up the scale (e.g. lifsat1, lifsat2, lifsat3, lifsat4, lifsat5). Move these into the
box marked Items.

3. In the Model section, make sure Alpha is selected.

4. In the Scale label box, type in the name of the scale or subscale (Life Satisfaction).

5. Click on the Statistics button. In the Descriptives for section, select Item, Scale, and Scale if item deleted. In the Inter-
Item section, click on Correlations. In the Summaries section, click on Correlations.

6. Click on Continue and then OK

Choosing the right statistic

Exploring relationships

• Eg. Correlation, Multiple regression

Exploring differences between groups

• Eg. T-tests, One-way ANOVA

Exploring relationship
Parametric Non IV DV
Parametric
Chi-Square test Categorical Categorical
Correlation Pearson’s Spearman’s Two
Correlation rank order continuous
correlation variables
Prediction Multiple None Continuous Continuous
regression
Prediction Logistic None Continuous Categorical (2
regression categories)
Differences between Groups (Independent
sample)
Parametric Non- parametric IV DV
None Chi-Square Categorical Categorical
Difference Independent Mann-Whitney U Categorical Continuous
between two sample T-test test
groups
Difference One-way ANOVA Kruskal Wallis Categorical Continuous
between two or
more than two
groups
Differences between Groups (Paired Sample)
Parametric Non- parametric IV DV
None Chi-Square Categorical Categorical
Compare time 1 Paired sample T- Wilcoxon signed Categorical Continuous
to time 2 test rank test test
Difference Repeated Friedman test Categorical Continuous
between two or measures ANOVA
more than two
groups
Pearson’s Correlation

Example of research question: Is there a relationship between the amount of control

people have over their internal states and their levels of perceived stress? Do people with
high levels of perceived control experience lower levels of perceived stress?

What you need: Two variables: both continuous, or one continuous and the other
dichotomous (two values).

What it does: Correlation describes the relationship between two continuous variables, in
terms of both the strength of the relationship and the direction.
Non-parametric alternative: Spearman Rank Order Correlation (rho).
Assumptions
• Both IV should be continuous
• Normality
• Outliers
• Linearity
1. From the menu at the top of the screen, click on Analyze, then select Correlate, then

2. Bivariate.

3. Select your two variables and move them into the box marked Variables (e.g. Total perceived
stress: tpstress, Total PCOISS: tpcoiss). If you wish you can list a whole range of variables here,
not just two.

4. In the Correlation Coefficients section, the Pearson box is the default option. If you wish to
request the Spearman rho (the non-parametric alternative), tick the Spearman box instead.

5. Click on the Options button. For Missing Values, click on the Exclude cases pairwise box. Under
Options, you can also obtain means and standard deviations if you wish.

6. Click on Continue and then on OK

small r=.10 to .29

medium r=.30 to .49

large r=.50 to 1.0

Negative sign refers only to the direction of the relationship, not the
strength.
Presenting the results

The relationship between perceived control of internal states (as measured

by the PCOISS) and perceived stress (as measured by the Perceived Stress
Scale) was investigated using Pearson product-moment correlation
coefficient. Preliminary analyses were performed to ensure no violation of
the assumptions. There was a strong, negative correlation between the
two variables, r = –.58, n = 426, p < .001, with high levels of perceived
control associated with lower levels of perceived stress.
Multiple Correlation

Research questions:

• How well a set of variables is able to predict a particular outcome

• Which variable in a set of variables is the best predictor of an outcome

• Whether a particular predictor variable is still able to predict an

outcome when the effects of another variable are controlled for (e.G.
Socially desirable responding).
Assumptions of Multiple Regression
Sample size: 15 participants per predictor are needed for a reliable equation

Multicollinearity and singularity: independent variables are highly correlated (r=.9 and above).

Outliers

Normality, linearity, homoscedasticity, independence of residuals

• normality: the residuals should be normally distributed about the predicted DV scores

• linearity: the residuals should have a straight-line relationship with predicted DV scores

• homoscedasticity: the variance of the residuals about predicted DV scores should be the same for all
predicted scores.
1.From the menu at the top of the screen, click on Analyze, then select Regression, then
Linear.

2.Click on your continuous dependent variable (e.g. Total perceived stress: tpstress) and
move it into the Dependent box.

3.Click on your independent variables (Total Mastery: tmast; Total PCOISS: tpcoiss) and click
on the arrow to move them into the Independent(s) box.

4.For Method, make sure Enter is selected. (This will give you standard multiple regression.)

5.Click on the Statistics button.

• Select the following: Estimates, Confidence Intervals, Model fit, Descriptives, Part and
partial correlations and Collinearity diagnostics.
• In the Residuals section, select Casewise diagnostics and Outliers outside standard deviations. Click on
Continue.

6. Click on the Options button. In the Missing Values section, select Exclude cases pairwise. Click on
Continue.

7. Click on the Plots button.

• Click on *ZRESID and the arrow button to move this into the Y box.
• Click on *ZPRED and the arrow button to move this into the X box.
• In the section headed Standardized Residual Plots, tick the Normal probability plot option.

Click on Continue.

8. Click on the Save button.

• In the section labelled Distances, select and Cook’s.

• Click on Continue and then OK.

INTERPRETATION OF OUTPUT FROM
STANDARD MULTIPLE REGRESSION
Step 1: Checking the assumptions

Multicollinearity

table labelled Correlations (r more than 0.3 and less than 0.8)

table labelled Coefficients (VIF less than 10)

Outliers, normality, linearity, homoscedasticity, independence of residuals

In the Normal P-P Plot, you are hoping that your points will lie in a reasonably straight diagonal line from bottom left to top
right.

In the Scatterplot of the standardised residuals (the second plot displayed) you are hoping that the residuals will be roughly
rectangularly distributed, with most of the scores concentrated in the centre

Outliers can be checked by inspecting the Cook’s distance

Step 2: Evaluating the model

To assess the statistical significance of the result, it is necessary to look

in the table labelled ANOVA. This tests the null hypothesis that multiple
R in the population equals 0

R Square and Adjusted R Square

Step 3: Evaluating each of the independent variables

Presenting the results of Multiple regression

Multiple regression was used to assess the ability of two control measures (Mastery
Scale, Perceived Control of Internal States Scale: PCOISS) to predict levels of stress
(Perceived Stress Scale). Preliminary analyses were conducted to ensure no violation of
the assumptions of normality, linearity, multicollinearity and homoscedasticity. The
model was significant (p <0.001) and the value of adjusted R square was 0.457.

In the final model, both the variables were statistically significant, with the Mastery
Scale recording a higher beta value (beta = –.44, p < .001) than the PCOISS Scale (beta =
–.33, p < .001).
Research Problem
Example of Broad Problem areas:

Staff turnover is higher than anticipated

Suppliers frequently deliver products with inconsistent quality, leading

to production delays and potential quality issues in finished goods.

Insufficient transportation capacity results in difficulties in meeting

customer demand and potential delays in delivery schedules.
Problem vs Symptoms
How to overcome the issue: “Staff turnover is higher than anticipated”

5 Whys

My best employees are leaving the organization. Why?

Because they are not satisfied with their jobs. Why?

Because they do not find a challenge in their jobs. Why?

Because they do not have control over their work. Why?

Because they do not have a lot of influence over planning, executing, and evaluating the work they do.
Why?

Because we have been reluctant to delegate.

Defining the problem statement

A good problem statement includes both a statement of the research

objective(s) and the research question(s).
Examples of research objectives

To find out what motivates consumers to buy a product online.

To analyze the impact of different supplier relationship management practices

on supply chain performance.

To study the adoption and integration of emerging technologies (e.g., IoT,

blockchain, AI) in supply chain operations.

To develop criteria for effective supplier selection, considering factors such as

quality, lead time, cost, and sustainability practices.
 A Comparison of Basic Research Designs
Blank Exploratory Descriptive Causal
Objective: Discovery of ideas and insights Describe market characteristics or Determine cause-and-effect
functions relationships
Characteristics: Flexible, versatile Marked by the prior formulation Manipulation of one or
of specific hypotheses more independent variables

Often the front end of total Preplanned and structured design Measure the effect on
research design dependent variable(s)
Control of other mediating
variables
Methods: Expert surveys Secondary data: quantitative Experiments
Pilot surveys analysis
Case studies Surveys
Secondary data: qualitative Panels
analysis Observation and other data
Qualitative research
Cross-Sectional Designs
• Involve the collection of information from any given sample of
population elements only once.
• In single cross-sectional designs, there is only one sample of
respondents and information is obtained from this sample only once.
• In multiple cross-sectional designs, there are two or more samples of
respondents, and information from each sample is obtained only once.
Often, information from different samples is obtained at different times.
• Cohort analysis consists of a series of surveys conducted at appropriate
time intervals, where the cohort serves as the basic unit of analysis. A
cohort is a group of respondents who experience the same event within
the same time interval.
Longitudinal Designs

A fixed sample (or samples) of population elements is measured

repeatedly on the same variables

A longitudinal design differs from a cross-sectional design in that the

sample or samples remain the same over time
Potential Sources of Error in Research
Designs
The Sampling Design Process
Define the Target Population (1 of 2)
The target population is the collection of elements or objects that possess the information
sought by the researcher and about which inferences are to be made. The target population
should be defined in terms of elements, sampling units, extent, and time.

• An element is the object about which or from which the information is desired, e.g., the respondent.
• A sampling unit is an element, or a unit containing the element, that is available for selection at some
stage of the sampling process.
• Extent refers to the geographical boundaries.
• Time is the time period under consideration.
Classification of Sampling Techniques
Convenience Sampling
Convenience sampling attempts to obtain a sample of convenient elements.
Often, respondents are selected because they happen to be in the right place at
the right time.
• Use of students, and members of social organizations

• Mall intercept interviews without qualifying the respondents

Judgmental Sampling

Judgmental sampling is a form of convenience sampling in which the

population elements are selected based on the judgment of the
researcher.

• Test markets

• Purchase engineers selected in industrial marketing research

• Expert witnesses used in court

Quota Sampling
Quota sampling may be viewed as two-stage restricted judgmental sampling.
• The first stage consists of developing control categories, or quotas, of population elements.

• In the second stage, sample elements are selected based on convenience or judgment.

Blank Population Composition Sample Composition Blank

Control Characteristic Percentage Percentage Number

Sex Blank Blank Blank

Male 48 48 480

Female 52 52 520

Blank 100 100 1,000

Snowball Sampling

In snowball sampling, an initial group of respondents is selected, usually at random.

• After being interviewed, these respondents are asked to identify others who
belong to the target population of interest.
• Subsequent respondents are selected based on the referrals.
Simple Random Sampling

Each element in the population has a known and equal probability of

selection.

Each possible sample of a given size (n) has a known and equal
probability of being the sample actually selected.

This implies that every element is selected independently of every

other element.
Systematic Sampling

The sample is chosen by selecting a random starting point and then picking
every ith element in succession from the sampling frame.

The sampling interval, i, is determined by dividing the population size N by

the sample size n and rounding to the nearest integer.

When the ordering of the elements is related to the characteristic of interest,

systematic sampling increases the representativeness of the sample.
Stratified Sampling

A two-step process in which the population is partitioned into subpopulations, or

strata.

The strata should be mutually exclusive and collectively exhaustive in that every
population element should be assigned to one and only one stratum and no
population elements should be omitted.

Next, elements are selected from each stratum by a random procedure, usually SRS.

A major objective of stratified sampling is to increase precision without increasing

cost.
Stratified Sampling

The elements within a stratum should be as homogeneous as possible, but

the elements in different strata should be as heterogeneous as possible.

The stratification variables should also be closely related to the

characteristic of interest.

Finally, the variables should decrease the cost of the stratification process
by being easy to measure and apply.
Cluster Sampling

The target population is first divided into mutually exclusive and collectively
exhaustive subpopulations, or clusters.

Then a random sample of clusters is selected, based on a probability

sampling technique such as SRS.

For each selected cluster, either all the elements are included in the sample
(one-stage) or a sample of elements is drawn probabilistically (two-stage).
Cluster Sampling

Elements within a cluster should be as heterogeneous as possible, but

clusters themselves should be as homogeneous as possible. Ideally,
each cluster should be a small-scale representation of the population.
Cluster Sampling Vs. Stratified Sampling
Table 11.3 Differences Between Stratified and Clustering
Sampling
Factor Stratified Sampling Cluster Sampling (One-Stage)
Objective Increase precision Decrease cost
Subpopulations All strata are included A sample of clusters is chosen
Within subpopulations Each stratum should be Each cluster should be
homogeneous heterogeneous
Across subpopulations Strata should be heterogeneous Clusters should be homogeneous
Sampling frame Needed for the entire Needed only for the selected
population clusters
Selection of elements Elements selected from each All elements from each selected
stratum randomly cluster are included
Calculating Sample Size (Small to moderate)

• n= sample size
• N= Number of people in the population
• p= Estimated variance for the population of area
• A= Desired precision expressed in decimal in the formula
• z= Required confidence level
Calculating Sample Size (Small to moderate)

• Let population size is 1000

Sample Size (Very large population)
• s=
Sample Size (Very large population)
• s=
Type 1 and Type 2 errors
A Classification of Scaling Techniques
Figure 8.2 A Classification of Scaling Techniques
A Comparison of Scaling Techniques

• Comparative scales involve the direct comparison of stimulus objects.

Comparative scale data must be interpreted in relative terms and
have only ordinal or rank order properties.

• In noncomparative scales, each object is scaled independently of the

others in the stimulus set. The resulting data are generally assumed to
be interval or ratio scaled.
Relative Advantages of Comparative Scales
• Small differences between stimulus objects can be detected.
• Same known reference points for all respondents.
• Easily understood and can be applied.
• Involve fewer theoretical assumptions.
• Tend to reduce halo or carryover effects from one judgment to
another.
Relative Disadvantages of Comparative
Scales
• Ordinal nature of the data
• Inability to generalize beyond the stimulus objects scaled
Comparative Scaling Techniques: Paired
Comparison Scaling
• A respondent is presented with two objects and asked to select one
according to some criterion.
• The data obtained are ordinal in nature.
• Paired comparison scaling is the most widely used comparative
scaling technique.
• With n brands, [n(n − 1) /2] paired comparisons are required.
• Under the assumption of transitivity, it is possible to convert paired
comparison data to a rank order.
Obtaining Shampoo Preferences Using Paired
Comparisons
Figure 8.3 Obtaining Shampoo Preferences Using Paired
Comparisons
Noncomparative Scaling Techniques

• Respondents evaluate only one object at a time, and for this reason
non-comparative scales are often referred to as monadic scales.

• Non-comparative techniques consist of continuous and itemized

rating scales.
Continuous Rating Scale (1 of 2)
Respondents rate the objects by placing a mark at the appropriate position on a line that
runs from one extreme of the criterion variable to the other.
The form of the continuous scale may vary considerably.
How would you rate Wal-Mart as a department store?
Version 1
Probably the worst - - - - - - I - - - - - - - - - - - - - Probably the best
Version 2
Probably the worst - - - - - - I - - - - - - - - - - - - -Probably the best
0 10 20 30 40 50 60 70 80 90 100
Continuous Rating Scale (2 of 2)
Version 3
Very bad Neither good nor bad Very good
Probably the worst - - - - - - - I - - - - - - - - - - - - -Probably the best
0 10 20 30 40 50 60 70 80 90 100
Itemized Rating Scales

• The respondents are provided with a scale that has a number or brief
description associated with each category.

• The categories are ordered in terms of scale position, and the respondents
are required to select the specified category that best describes the object
being rated.

• The commonly used itemized rating scales are the Likert, semantic
differential, and Stapel scales.
Likert Scale
The Likert scale requires the respondents to indicate a degree of
agreement or disagreement with each of a series of statements about
the stimulus objects.
Blank Strongly Disagree Neither agree Agree Strongly
disagree nor disagree agree
1. Wal-Mart sells high-quality merchandise. 1 2X 3 4 5
2. Wal-Mart has poor in-store service. 1 2X 3 4 5
3. I like to shop at Wal-Mart. 1 2 3X 4 5

• The analysis can be conducted on an item-by-item basis (profile

analysis), or a total (summated) score can be calculated.
• When arriving at a total score, the categories assigned to the
negative statements by the respondents should be scored by
reversing the scale.
Semantic Differential Scale
The semantic differential is a seven-point rating scale with end points
associated with bipolar labels that have semantic meaning.
WAL-MART IS:
Powerful --:--:--:--:-X-:--:--: Weak
Unreliable --:--:--:--:--:-X-:--: Reliable
Modern --:--:--:--:--:--:-X-: Old-fashioned
The negative adjective or phrase sometimes appears at the left side of the
scale and sometimes at the right.
This controls the tendency of some respondents, particularly those with very
positive or very negative attitudes, to mark the right- or left-hand sides
without reading the labels.
Individual items on a semantic differential scale may be scored on either a −3
to +3 or a 1 to 7 scale.
A Semantic Differential Scale for Measuring Self- Concepts,
Person Concepts, and Product Concepts
1) Rugged :---:---:---:---:---:---:---: Delicate
2) Excitable :---:---:---:---:---:---:---: Calm
3) Uncomfortable :---:---:---:---:---:---:---: Comfortable
4) Dominating :---:---:---:---:---:---:---: Submissive
5) Thrifty :---:---:---:---:---:---:---: Indulgent
6) Pleasant :---:---:---:---:---:---:---: Unpleasant
7) Contemporary :---:---:---:---:---:---:---: Obsolete
8) Organized :---:---:---:---:---:---:---: Unorganized
9) Rational :---:---:---:---:---:---:---: Emotional
10) Youthful :---:---:---:---:---:---:---: Mature
11) Formal :---:---:---:---:---:---:---: Informal
12) Orthodox :---:---:---:---:---:---:---: Liberal
13) Complex :---:---:---:---:---:---:---: Simple
14) Colorless :---:---:---:---:---:---:---: Colorful
15) Modest :---:---:---:---:---:---:---: Vain
Stapel Scale
The Stapel scale is a unipolar rating scale with ten categories
numbered from -5 to +5, without a neutral point (zero). This scale
is usually presented vertically.
Wal-Mart
+5 +5
+4 +4
+3 +3
+2 +2X
+1 +1
HIGH QUALITY POOR SERVICE
-1 -1
-2 -2
-3 -3
-4X -4
-5 -5

The data obtained by using a Stapel scale can be analyzed in the same
way as semantic differential data.
Basic Noncomparative Scales
Table 9.1 Basic Noncomparative Scales
Scale Basic Characteristics Examples Advantages Disadvantages
Continuous rating Place a mark on a Reaction to TV Easy to construct Scoring can be
scale continuous line commercials cumbersome unless
computerized

Itemized Rating Blank Blank Blank Blank

Scales

Likert scale Degree of agreement on a Measurement of Easy to construct, More time consuming
1 (strongly disagree) to 5 attitudes administer, and
(strongly agree) scale understand

Semantic Seven-point scale with Brand, product, and Versatile Controversy as to

differential bipolar labels company images whether the data are
interval

Stapel scale Unipolar ten-point scale, 25 Measurement of Easy to construct; Confusing and difficult
to 15, without a neutral attitudes and images administered over to apply
point (zero) telephone
Summary of Itemized Scale Decisions
Table 9.2 Summary of Itemized Rating Scale Decisions
1. Number of categories Although there is no single, optimal number, traditional guidelines
suggest that there should be between five and nine categories.

2. Balanced versus unbalanced In general, the scale should be balanced to obtain objective data.

3. Odd or even number of categories If a neutral or indifferent scale response is possible from at least
some of the respondents, an odd number of categories should be
used.
4. Forced versus nonforced In situations where the respondents are expected to have no
opinion, the accuracy of data may be improved by a nonforced
scale.
5. Verbal description An argument can be made for labeling all or many scale categories.
The category descriptions should be located as close to the
response categories as possible.

6. Physical form A number of options should be tried and the best one selected.
Balanced and Unbalanced Scales
Figure 9.1 Balanced and Unbalanced Scales
Rating Scale Configurations
Figure 9.2 Rating
Scale Configurations
Some Unique Rating Scale Configurations
Figure 9.3 Some
Unique Rating Chart
Configurations
Some Commonly Used Scales in Marketing
Table 9.3 Some Commonly Used Scales in Marketing

Construct Blank Blank Scale Descriptors Blank Blank

Attitude Very bad Bad Neither bad nor good Good Very good

Importance Not at all Not important Neutral Important Very important

important
Satisfaction Very Dissatisfied Neither dissatisfied nor Satisfied Very satisfied
dissatisfied satisfied
Purchase Definitely will Probably will Might or might not buy Probably Definitely will
Intent not buy not buy will buy buy
Purchase Never Rarely Sometimes Often Very often
Frequency
Assumptions of Non-parametric statistics

• Random samples.

• Independent observations.
Chi-square test for independence
• Explore the relationship between two categorical variables.

• When a 2 by 2 table (two categories in each variable), the output from chi-square includes an
additional correction value (yates’ correction for continuity).

• Additional assumptions: the lowest expected frequency in any cell should be 5 or more.
Some authors suggest less stringent criteria: at least 80 per cent of cells should have expected
frequencies of 5 or more. If you have a 2 by 2 table, it is recommended that the expected
frequency be at least 10. If you have a 2 by 2 table that violates this assumption, you should
consider reporting fisher’s exact probability test instead.
1. From the menu at the top of the screen, click on Analyze, then Descriptive Statistics, and then Crosstabs.

2. Click on one of your variables (e.g. sex) to be your row variable and click on the arrow to move it into the
box marked Row(s).

3. Click on the other variable to be your column variable (e.g. smoker) and click on the arrow to move it into
the box marked Column(s).

4. Click on the Statistics button. Tick Chi-square and Phi and Cramer’s V. Click on Continue.

5. Click on the Cells button.

6. • In the Counts box, make sure there is a tick for Observed.

7. • In the Percentage section, click on the Row box.

• In the Residuals section click on Adjusted standardized.

8. Click on Continue and then OK

Effect size
phi coefficient (2x2)

.10 for small effect, .30 for medium effect and .50 for large effect.

Larger than (2x2)

Cramer’s V

For R–1 or C–1 equal to 1 (two categories): small=.01, medium=.30, large=.50

For either R–1 or C–1 equal to 2 (three categories): small=.07, medium=.21, large=.35

For either R–1 or C–1 equal to 3 (four categories): small=.06, medium=.17, large=.29
Reporting

A chi-square test for independence (with Yates’ Continuity Correction)

indicated no significant association between gender and smoking
status, χ2 (1, n = 436) = .34, p = .56, phi = –.03.
MANN-WHITNEY U TEST
1. From the menu at the top of the screen click on Analyze, then Nonparametric Tests and then on
Independent Samples.

2. In the Objective tab click on Customize analysis in the section “What is your objective?”.

3. Click on the Fields tab.

4. Click on your categorical (independent) variable (e.g. Sex) and move it into the Groups box.

5. Click on your continuous (dependent) variable (e.g. Total Self esteem: tslfest) and move it into the Test
Fields box.

6. Click on the Settings tab and select Customize tests. Click on Mann-Whitney U (2 samples).

7. Click on the Run button

Median Values
1. From the menu at the top of the screen, click on Analyze, then select Compare means and choose Means.

2. Click on your continuous variable (e.g. total self-esteem: tslfest) and move it into the Dependent List box.

3. Click on your categorical variable (e.g. sex) and move it into the Independent List box.

4. Click on the Options button. Click on Median in the Statistics section and move into the Cell

5. Statistics box. Click on Mean and Standard Deviation and remove from the Cell Statistics box.

6. Click on Continue.

7. Click on OK
WILCOXON SIGNED RANK TEST
1. From the menu at the top of the screen, click on Analyze, then select Nonparametric Tests,

then Legacy Dialogs and then 2 Related Samples.

2. Click on the variables that represent the scores at Time 1 and at Time 2 (e.g. fear of stats time1: fost1,

fear of stats time2: fost2). Click on the arrow to move these into the Test Pairs box.

3. Make sure that the Wilcoxon box is ticked in the Test Type section.

4. Click on the Options button. Choose Quartiles (this will provide the median scores for each

time point).

5. Click on Continue and then on OK