1.

SPSS tutorial exercise 1

GETTING TO KNOW SPSS

Objective: To introduce you to the SPSS operating environment. After this exercise you will
understand the following:
1 the Date Editor 4 the main Menu functions
2 the elements of the SPSS spreadsheet 5 the use of online SPSS Help
3 the role of ‘variables’ and ‘cases’

1. Start SPSS
Click Start, Programs, SPSS for Windows, and SPSS 25 or 26 for Windows. When SPSS is
running, select Type in data, OK.
2. Menus
Briefly explore the menus at the top of the screen. As there is currently no data in SPSS, few
functions in the menu are currently active.

File – to create, open, save and read files.

Edit – to modify or copy text.

View – to change characteristics of the window.

Data – to define variables and make global changes to data files

Transform – to change variables or create new variables.

Analyze – to select a statistical procedure.
Graphs – to create a variety of graph types from the data.

3. Data Editor
The Data Editor window is the empty spreadsheet in which data from your marketing research
project can be added. The columns in the spreadsheet denoted ‘var’ at present are for the
‘variables’ in your research project. Each numbered row represents each ‘case’ in your research.
A case is often a respondent if your data has been gathered from a survey or questionnaire, but
a case can also relate to other things such as year, customer, geographic region, or sales person,
for example. The tab labelled Data View at the bottom of the spreadsheet indicates that we are
currently viewing data. Now click on the adjacent Variable View tab. Now you will see that the
columns have changed. These columns indicate different attributes for each variable. We will
explore each of these attributes.
4. Open a Data File
Click File, Open, and Data.

Select the HobbitData file. Double-click on HobbitData.sav. The spreadsheet will now fill with
the data, variables and variable characteristics for this particular case. In the remaining time,
explore this data.
5. Explore SPSS Help
Click in Help on the menu. You will see that you can find help on any Topic, learn SPSS by
running tutorials, exploring case studies, or using help from the statistics coach. The main source
of help is accessed by selecting Topics. This operates like a User Manual with a table of contents
(Contents), an Index, as well as being able to Search for specific information or help on a
particular topic. Briefly familiarize yourself with this help function. You are certain to want to use
it later in the tutorial for this Subject.
Save the Output as: TutorialOne_Firstname_Lastname

Then submit to LMS Via the link

Reference : Coakes and Steed (2011), Chapter 1
2. SPSS tutorial exercise 2
PREPARING SPSS FILES AND ENTERING DATA

Objective: To show you how to set up the SPSS spreadsheet (i.e. data and variable views)
so they are ready to enter data from a particular survey. You will learn how to:

1. define each of the variables from your survey

2. enter the data into the Data View

Attached are 3 examples of responses from a part of a questionnaire which Professor

Crouch conducted recently investigating how convention meeting planners evaluate and
choose convention sites. This part of the survey deals with some questions which were
concerned with the characteristics of the survey respondents and the association they
worked for.

Do the following:

1. Click on the Variable View Tab

There are 13 questions but 17 variables in this section of the survey since each of questions
6, 9, 10, and 13 consist of two parts. In the Variable View, each row is used to define one
variable at a time.
2. Define a Name for the first variable
The first column is labelled Name. Type in a unique name to signify the first variable in the
survey, to a maximum of 8 characters.
3. Define the variable Type
In the next column (Type), click on the cell and then click on the shaded area with the 3 dots.
This opens a box indicating 8 types of variables. Select the type that corresponds to the first
variable. Use the Help button to read what each variable type means. If you are dealing with
a real number, you should also indicate the maximum number of digits in the data for this
variable (including the decimal point), and the number of decimal places you require. When
you have made your choice, click OK.
4. Specify Width and Decimals
The width and decimal values you entered (if any) will then automatically appear in the Width
and Decimals columns. If not, type these in.

5. Label the variable

In the Label column, you should type in a more detailed description of the variable (up to 256
characters but it is better to keep it brief) that explains it in more detail than is possible with
the Name which must be limited to 8 characters.
6. Specify Values for a categorical variable
Where the variable deals with data that is really only just a proxy for non-numeric data, the
values use to represent these proxies can be defined in the Values column. Click on the
shaded area with the 3 dots to open the Value Labels box. For each value you can indicate
a relevant label and create a list of labels for all values.
7. Specify Missing data
If, in your survey, you wish to define which codes representing various ‘missing data’ (such
as ‘question not completed’, ‘don’t know’, ‘not applicable’, etc.) such values can be defined
as missing data. SPSS recognizes these and drops them from any analysis involving this
particular variable.
8. Specify the number of Columns you require
Under Columns type the total width (in characters) you wish the variable column to have in
the Data View. Normally this would be no less than the value under Width.
9. Specify Alignment in the variable column
In the Align column, you have the option of aligning the data in the Data-View column as
either left-justified, right-justified, or centre-justified.

10. Specify the type of Measure the variable uses

Finally, in the Measure column you can specify what type of data you are dealing with for
this variable. The choices are Scale, Nominal, or Ordinal. Nominal data is data which is not
ordered or scaled. In other words, the data stands for some type, or category. Ordinal data is
data which has some order to it, such as a respondent’s ranking of 6 products from most-
preferred to least-preferred. Scaled data is data which is ordered but for which each step in
the order is comprised of equal intervals. Scaled data may be either interval or ratio data.
Ratio data has a true zero point such that ratios in the data are meaningful. If this is not the
case then scaled data is of the interval type. SPSS then uses this information when analyses
are performed on the variable.
11. Repeat 2 to 10 for all other variables
Repeat this procedure until all of the 17 variables in the survey have been defined.
12. Enter data
Return to the Data View and you are now ready to enter the data. You will notice that this
view now shows the variable name at the top of each column. Click on the first cell in the first
row of the spreadsheet and start to type in the data. Continue across the page until the first
survey is complete. Then continue the process until the first 3 rows of data have been
entered for the 3 survey examples.

13. Editing data or variables

To insert a new variable, select the Variable View, highlight the existing variable above
which you wish the new variable to appear.
Then select Data, Insert variable.

To insert a new row of data (called a ‘case’ in SPSS) in the Data View, highlight the case
above which you wish the new case to appear. Then select Data, Insert case
14. Entering data by cutting and pasting from other spreadsheets
Data can also be entered into SPSS if it already exists elsewhere in other applications such
as MS Excel. For example, you can highlight the cells in MS Excel, select Edit, Copy then
from the first cell in the SPSS Data View paste in the data by selecting Edit, Paste
Reference
Coakes and Steed (2011), Chapter 2
 MKT2MRE - SPSS Tutorial 3 – Step by Step

MKT2MRE - SPSS Tutorial 3 – Step by Step

1. Open the database MREexercise3.sav

2.

Page 1 of 23
 MKT2MRE - SPSS Tutorial 3 – Step by Step

Check for any errors in the data

Select the Analyze menu. Click on Descriptive Statistics and then Frequencies… to open the
Frequencies dialog box. Select hope1 to hope6 and click the right-arrow button to move these
variables into the Variable(s): box. Click OK to obtain the output. Examine this output. You will
see that there is some missing data (coded as 9) to indicate that some students did not answer
all questions. However, there is one error (due to mistyping at data entry). Note the variable
which contains the error

Page 2 of 23
MKT2MRE - SPSS Tutorial 3 – Step by Step

Page 3 of 23
 MKT2MRE - SPSS Tutorial 3 – Step by Step

3. Find and correct the error

To find where that error is in the database you can either scroll down the relevant column for
that variable until you find it or click on the variable column and then click Edit, Find… to open
the Find dialog box. In the Find what field, type the incorrect value you wish to find, and click
Find Next. Has it found the incorrect value? If not, click Find Next again until you are sure you
have found the error and type in the value you know to be correct (5) after checking the original
survey for that student.

Page 4 of 23
MKT2MRE - SPSS Tutorial 3 – Step by Step

Page 5 of 23
MKT2MRE - SPSS Tutorial 3 – Step by Step

Page 6 of 23
 MKT2MRE - SPSS Tutorial 3 – Step by Step

4. Recode negatively worded items

As we are going to compute an overall measure of how each student feels about their future, we
need to make sure that the scale for each of the 6 questions is consistent. Note that there are 2
negatively worded questions such that a 5 on the scale for these questions points in the
opposite direction from the other 4 questions. Which 2 questions are these? Select Transform,
Recode, Into Different Variables… to open the Recode into Different Variables dialog box.
Shift the first variable to be recoded into the Numeric Variable->Output Variable: box. In the
Name box, type a new name so that the old variable remains (e.g. hope3.1). Click Change.
Click Old and New Values… As we wish to change 5 to 1 and 1 to 5 etc., type 1 in the Old
Value, Value: field and type 5 in the New Value, Value: field. Then click Add. Repeat for the
other values to be recoded until the Old → New box is complete. Click Continue, OK. Check to
see that the recoding has occurred correctly. Repeat the same procedure to recode the other
variable.

Page 7 of 23
MKT2MRE - SPSS Tutorial 3 – Step by Step

Page 8 of 23
MKT2MRE - SPSS Tutorial 3 – Step by Step

Page 9 of 23
 MKT2MRE - SPSS Tutorial 3 – Step by Step

5.

Page 10 of 23
 MKT2MRE - SPSS Tutorial 3 – Step by Step

5. Recode missing values using Mean substitution

When missing data occurs, SPSS is clearly unable to use any cases in an analysis requiring the
use of missing data. The analyst has two options: a) accept that data is missing and remove
those cases affected from the analysis (which SPSS does automatically), or b) recode the
missing data, so that the cases can be used, by using an estimate. One suitable estimate might
be to substitute the mean for the missing data as follows. First we need to calculate the mean of
each variable. Select Analyze, Descriptive Statistics, Descriptives… Shift the 6 variables
(using the two that were transformed in 4 above) into the Variable(s): box. Click Options… and
check the Mean box, unchecking the other boxes. Click Continue, OK and write down the
calculated mean values. Select Transform, click Recode and Into Same Variables to open the
Recode into Same Variables dialog box. Select the variable you wish to recode (you only need
to do this for variables which have missing data) and move them to the Variables: box. Click
Old and New Values… In the Old Value box select System- or user-missing button. In the
New Value box type the mean for that variable and click Add, Continue, and OK. Repeat until
all variables have their missing data recoded in this way.

Page 11 of 23
MKT2MRE - SPSS Tutorial 3 – Step by Step

Page 12 of 23
MKT2MRE - SPSS Tutorial 3 – Step by Step

Page 13 of 23
 MKT2MRE - SPSS Tutorial 3 – Step by Step

REPEAT for each variable

Page 14 of 23
MKT2MRE - SPSS Tutorial 3 – Step by Step

Page 15 of 23
 MKT2MRE - SPSS Tutorial 3 – Step by Step

6. Compute a total hope score

Click Transform, Compute… In the Target Variable: box, type hopetot. Scroll through the
available numeric expressions until you find SUM(numexpr, numexpr,…), select it and click the
up arrow to move this function into the Numeric Expression: box. Inside the brackets of the
expression, type or shift in all 6 variables separated by commas to replace the question marks.
Then click OK and check the results.

Page 16 of 23
MKT2MRE - SPSS Tutorial 3 – Step by Step

Page 17 of 23
MKT2MRE - SPSS Tutorial 3 – Step by Step

Page 18 of 23
 MKT2MRE - SPSS Tutorial 3 – Step by Step

7. Screen the composite variable for normality

Select Analyze, Descriptive Statistics, and Explore… Select hopetot and shift it into the
Dependent List: box. Click Plots… Check Histogram, and Normality plots with tests boxes
and ensure that the Factor levels together button is selected in the Boxplots display. Click
Continue, ensure that Both is activated. Click OK and observe the results. Does the histogram
look like a normal distribution? If the Kolmogorov-Smirnov significance level is greater than
0.05, then normality is assumed.

Page 19 of 23
MKT2MRE - SPSS Tutorial 3 – Step by Step

Page 20 of 23
MKT2MRE - SPSS Tutorial 3 – Step by Step

Page 21 of 23
MKT2MRE - SPSS Tutorial 3 – Step by Step

Page 22 of 23
 MKT2MRE - SPSS Tutorial 3 – Step by Step

Save Output as “Your_Name_Tutorial_3” and Submit through LMS

Page 23 of 23
 MKT2MRE – SPSS TUTORIAL 4 – STEP BY STEP

MKT2MRE - SPSS Tutorial 4 – Step by Step

Objective: Having screened and cleaned the data, and having transformed the data in order to
obtain the necessary variables for analysis, you are now ready to begin analysis. The first step
of any analysis is to explore the data so that you can summarise, describe and understand what
you have collected through your research. This will give you a better ‘feel’ for the data. It will
help you understand the characteristics of the sample you are working with and, by comparing
these characteristics with any known characteristics of the population you had planned the
sample to represent, you will be able to judge the representativeness of the data. By knowing
more about the data you will also be in a better position to undertake appropriate advanced
statistical analyses which are suitable to the type of data you are working with. Hence, in this
exercise, you will learn how to:
1. produce measures of the ‘central tendency’ of the data
2. produce measures of the amount of variability in the data
3. produce measures and charts which show how the data is distributed
We will use an example in this exercise from the Hobbit’s Choice restaurant. In the example,
dollars spent per month in a restaurant, and likelihood of patronizing the new restaurant by 400
people were recorded. Your task is to obtain a frequency table and the appropriate chart and
descriptive statistics for each of these two variables in the data.

1. Open the database MREHobbitData.sav

Click File, Open, and Data. Next to the Look in: box, click on the down arrow button to find and
select the MREHobbitData.sav file. Double-click on that file to open the data. The data
spreadsheet has 2 columns; a totspent column which indicates how much is spent per month on
restaurants, plus a likely column. Familiarise yourself with the data in the Variable View.

Page 1 of 12
 MKT2MRE – SPSS TUTORIAL 4 – STEP BY STEP

2. Obtain a frequency table and measure of central tendency and variability for the
sales variable

Select the Analyze menu. Click on Descriptive Statistics and then Frequencies… to open
the Frequencies dialog box. Select totspent and click the right-arrow button to move this
variable into the Variable(s): box. Click on the Statistics… command button to open the
Frequencies: Statistics dialogue box. In the Percentile Values box, check the Quartiles
box. In the Central Tendency box, check the Mean, Median and Mode boxes. In the
Dispersion box, check the Std. deviation, Variance, Range, Minimum and Maximum
boxes. Click Continue. Click on the Charts… button to open the Frequencies: Charts
dialogue box. Click on the Histogram(s) radio button and check the With normal curve
box. Click Continue and then OK to obtain the output. Examine this output.

Page 2 of 12
MKT2MRE – SPSS TUTORIAL 4 – STEP BY STEP

Page 3 of 12
MKT2MRE – SPSS TUTORIAL 4 – STEP BY STEP

Page 4 of 12
MKT2MRE – SPSS TUTORIAL 4 – STEP BY STEP

Page 5 of 12
MKT2MRE – SPSS TUTORIAL 4 – STEP BY STEP

Page 6 of 12
 MKT2MRE – SPSS TUTORIAL 4 – STEP BY STEP

3. Obtain a frequency table and measure of central tendency for the salesperson
rank

Select the Analyze menu. Click on Descriptive Statistics and then Frequencies… to open the
Frequencies dialog box. Select likely and click the right-arrow button to move this variable into
the Variable(s): box. Click on the Statistics… command button to open the Frequencies:
Statistics dialogue box. In the Percentile Values box, check the Quartiles box. In the Central
Tendency box, check the Mode box. In the Dispersion box tick Minimum and Maximum. No
other boxes should remain checked since likely is an ordinal variable. Click Continue. Click on
the Charts… button to open the Frequencies: Charts dialogue box. Click on the Bar chart(s)
radio button. Click Continue and then OK to obtain the output. Examine this output.

Page 7 of 12
MKT2MRE – SPSS TUTORIAL 4 – STEP BY STEP

Page 8 of 12
MKT2MRE – SPSS TUTORIAL 4 – STEP BY STEP

Page 9 of 12
MKT2MRE – SPSS TUTORIAL 4 – STEP BY STEP

Page 10 of 12
MKT2MRE – SPSS TUTORIAL 4 – STEP BY STEP

Page 11 of 12
 MKT2MRE – SPSS TUTORIAL 4 – STEP BY STEP

Save this Output as “Your_Name_Tutorial_4” and Submit through LMS

Reference
Coakes and Steed (2011), Chapter 4
Burns and Bush (2014), Chapter 12

Page 12 of 12
 SPSS tutorial exercise 5 – Step by Step
TESTING HYPOTHESES

Objective: With respect to those who are ‘very likely’ to patronize The Hobbit’s Choice
Restaurant, Jeff believes that they will either ‘very strongly’ or ‘somewhat’ prefer each of the
following: (a) waitstaff with tuxedos, (b) unusual desserts, (c) large variety of entrees, (e)elegant
décor, and (f) jazz combo music. Does the survey support or refute Jeff’s hypotheses?

Do the following:

1. Open the database MREHobbitData.sav

Click File, Open, and Data. Next to the Look in: box, click on the down arrow button to find and
select the MREHobbitData.sav file. Double-click on that file to open the data.

Familiarise yourself with the data in the Variable View for the variables likely, waitstaf, dessert,
variety, unusual, elegant, and jazz.
2. Select the respondents who indicated they were ‘very likely’ to patronize The Hobbit’s
Choice

Click Data, Select Cases… ,

and click the radio button, If condition is satisfied. Then click If…. To open the Select Cases:
If dialogue box.

Now select the variable likely, click the = button and type 5 which is the code for ‘very likely’.
Then click Continue and OK. Notice in the Data View that the cases which do not satisfy the
condition that likely = 5 have been crossed out.
3. Run one-sample t-tests to test the hypotheses indicated above.
Click Analyze, Compare Means, and One-Sample T Tests…

Select each of the variables likely, waitstaf, dessert, variety, unusual, elegant, and jazz. Type 4
in the Test Value: box to test the hypotheses for the ‘somewhat’ prefer case, and click OK.
Observe and interpret the results.

Now repeat this step again but instead use a Test Value: of 5 for ‘very strongly’ prefer and
interpret the results. Which of Jeff’s hypotheses are supported?

Reference
Coakes and Steed (2011), Chapter 6
Burns and Bush (2014), Chapter 13
 SPSS tutorial exercise 6 – Step by Step

TESTING FOR DIFFERENCES

Objective: The objective of this exercise to learn how to use the Independent Samples T-Test
and Analysis of Variance (ANOVA) to investigate whether significant differences exist in a
sample between two or more sub-samples (e.g., to see if males respond to a survey question
differently to females). We will consider the following questions:

1. Jeff wonders if The Hobbit’s Choice Restaurant is more appealing to women than it is to
men or vice versa. Perform the proper analysis, interpret it, and answer Jeff’s question.
2. In general, upscale establishments are appealing to higher-income households, whereas
they are less appealing to lower-income households. Is this pattern the case for The
Hobbit’s Choice Restaurant?

Do the following:
1. Open the database MREHobbitData.sav
Click File, Open, and Data. Next to the Look in: box, click on the down arrow button to find and
select the MREHobbitData.sav file. Double-click on that file to open the data.
Familiarise yourself with the data in the Variable View for the variables gender, likely, and
income.

2. Analyse question 1
Click Analyze, Compare Means, and Independent Samples T Test…
Select likely as the Test Variable and gender as the Grouping Variable.

Click Define Groups… and enter 1 (for male) in the Group1: field, and 2 (for female) in the
Group 2: field.

Then click Continue and OK. Interpret the results.

What is the answer to Jeff’s first question above?
3. Analyse question 2
Click Analyze, Compare Means, and One-Way ANOVA… to open the One-Way ANOVA
dialogue box.

Select likely to the Dependent List: and income as the Factor:. Click Post Hoc… to open the
One-Way ANOVA: Post Hoc Multiple Comparisons dialogue box.
Check the Duncan check box

and click Continue and OK.

Interpret the results. What is the answer to Jeff’s second question above?

Reference
Coakes and Steed (2011), Chapter 6 & 7
Burns and Bush (2014), Chapter 13
4. SPSS tutorial exercise 7 – Step by Step
TESTING ASSOCIATIONS

Objective: The objective of this exercise is to learn how to use Chi-Square Analysis in a cross-
tabulation to see if changes in one variable appear to be associated with changes in another
variable. For example, a consumer’s gender might be related to their price sensitivity for a
haircut. A further method for testing associations between variables involves the calculation of a
correlation coefficient which measures the strength of any association. We will consider the
following questions:

1. Which television news program would reach probable patrons of The Hobbit’s Choice
Restaurant most effectively?
2. Do older or younger people want unusual desserts and/or unusual entrees at The
Hobbit’s Choice Restaurant?

Do the following:

1. Open the database MREHobbitData.sav

Click File, Open, and Data. Next to the Look in: box, click on the down arrow button to find and
select the MREHobbitData.sav file. Double-click on that file to open the data.
Familiarise yourself with the data in the Variable View for the variables timenew, hcpatron,
yrborn, age, unusal and dessert.

2. Analyse question 1
Click Analyze, Descriptive Statistics, Crosstabs… to open the Crosstabs dialogue box.
Transfer the timenew variable into the Row(s): box and the hcpatron variable into the
Column(s): box. Next click the Cells… button to open the Crosstabs: Cell Display box

Check the Observed, Expected, Row, Column and Total check boxes. Then click Continue.
Now click the Statistics… button to open the Crosstabs: Statistics box.

Check the Chi-square box and click Continue.

Finally click the OK button to perform the analysis.

Interpret the results and answer question 1 above?

3. Analyse question 2
Click Analyze, Correlate, and Bivariate… to open the Bivariate Correlations dialogue box.

Select yrborn, age, unusal and dessert to the Variables: box.

Next check the Pearson check box and press the Two-tailed radio button as we are not
hypothesizing a directional relationship between age and menu preference.

Finally, click OK to calculate the table of correlations.

Interpret the results. What is the answer to the second question above?

Save as YOURNAME_TUTORIAL7 and Submit via LMS

Reference
Coakes and Steed (2011), Chapter 5
Burns and Bush (2014), Chapter 14
 SPSS tutorial exercise 8 – Step by Step
PREDICTIVE ANALYSIS USING MULTIPLE REGRESSION

Objective: The objective of this exercise is to learn how to use Multiple Regression Analysis in
order to investigate the linear relationship between one variable (commonly referred to as either
the dependent, endogenous or criterion variable) and one or more other variables (respectively
commonly referred to as either the independent, exogenous or predictor variables). It is
important to emphasize that regression analysis assumes a linear relationship as follows:

D = β0 + β1 . I 1 + β2 . I 2 + … + β n . I n + ε
where, D = dependent variable
I1, I2, …, In = independent variables,
β0, β1, β2, … ,βn = regression coefficients to be estimated,
n = number of independent variables,
ε = residual error (i.e. variation from a straight line relationship).

Regression analysis is not suited to analysing nonlinear relationships (unless they can be
transformed to a linear form).

Regression analysis makes several assumptions, as follows, the first of these being the
assumption of linearity:

1. linear relationship between dependent variable and each of the independent variables.
2. the residual errors (ε) are normally distributed.
3. the residual errors are independent and have equal variance.

As a minimum, there needs to be data points or cases exceeding five times the number of
independent variables (n).

We will consider the following question:

1. Develop a general conceptual model of market segmentation for the Hobbit’s Choice
Restaurant. Test it using multiple regression analysis and interpret your findings.

Do the following:

1. Open the database MREHobbitData.sav

Click File, Open, and Data. Next to the Look in: box, click on the down arrow button to find and
select the MREHobbitData.sav file. Double-click on that file to open the data. There are three
classes of variables the model could be based on: (1) demographic factors, (2) restaurant
patronage, and (3) restaurant feature preferences. The media usage variables are nominally
scaled and are therefore not suitable for regression analysis.
Familiarise yourself with the data in the Variable View for the variables:

Dependent variable – likely

Independent variables –
Demographic variables – yrborn (or age), reincome, famsize
Restaurant patronage – avprice, totspent
Restaurant feature preferences – water, drive, waitstaf, dessert,
variety, unusal, simple, elegant, string, jazz

2. Analyse question 1
Click Analyze, Regression, Linear… to open the Linear regression dialogue box.
Transfer the likely variable into the Dependent: box and the other independent variables into
the Independent(s): box. Next select Stepwise in the Method: box.
Click on the Statistics… button to open the Linear Regression: Statistics dialogue box and
check the boxes for Estimates, Model fit, and Collinearity diagnostics. Then click Continue.

Now click the Options… button to open the Liner Regression: Options dialogue box. Make
sure that the Use probability of F radio button is selected and ensure that the default values of
F for Entry: and Removal: of variables are set at 0.05 and 0.10 respectively. Check the
Include constant in the equation box if it is not already checked and click Continue and OK
to perform the analysis.
Interpret the results and answer question 1 above?

Reference
Coakes and Steed (2011), Chapter 17
Burns and Bush (2014), Chapter 15
 SPSS tutorial exercise 9 – STEP BY STEP
RELIABILITY ANALYSIS

Objective: In marketing, one is often faced with the task of trying to measure abstract
constructs such as ‘satisfaction’, ‘risk-averse attitudes’, ‘service quality’, or ‘donor motivation’,
etc. In these instances, such constructs are typically multifaceted in nature. That is, they can not
be adequately measured using a single item and thus a multi-item scale is typically developed
consisting of topics which span all or most of the facets of the construct in question. Collectively,
then, these items provide a more complete measure of the construct, but the suitability and
reliability of the scale depends on how well each item comprising the scale complements the
other items in the scale. In other words, do they seem to be measuring the same thing.
Cronbach’s alpha provides a measure of the internal consistency of a multi-item scale. It can
be interpreted as a correlation coefficient and it ranges in value from 0 to 1. The generally
agreed lower limit for an acceptable Cronbach’s alpha is 0.7. In this exercise you will learn how
to use SPSS to calculate Cronbach’s alpha.

We will consider the following question:

A researcher wished to determine the reliability of an 8-item, ‘optimism’ scale. Items 1, 3, 4 and
7 of the scale were negatively worded. A 5-point Likert scale response format was used for each
item. There were 363 survey responses (although there are some missing values). Determine
the Cronbach’s Alpha coefficient for the scale and make recommendations about the possible
exclusion of items.

Do the following:

1. Open the database MREexercise9.sav

Click File, Open, and Data. Next to the Look in: box, click on the down arrow button to find and
select the MREexercise9.sav file. Double-click on that file to open the data.

Familiarise yourself with the data in the Variable View for the variables opt1 to opt8.
2. Recode negatively worded items
We need to make sure that the scale for each of the 8 items is consistent. Note that there are 4
negatively worded items (opt1, opt3, opt4, and opt7) in the scale such that a 5 on the scale for
these questions points in the opposite direction from the other 4 items. Select Transform,
Recode, Into Different Variables… to open the Recode into Different Variables dialog box.
Shift the first variable to be recoded into the Numeric Variable->Output Variable: box. In the
Name box, type a new name so that the old variable remains (e.g. opt1.1). Click Change.

Click Old and New Values… As we wish to change 5 to 1 and 1 to 5 etc., type 1 in the Old
Value, Value: field and type 5 in the New Value, Value: field. Then click Add. Repeat for the
other values to be recoded until the Old → New box is complete. Click Continue, OK.
Check to see that the recoding has occurred correctly. Repeat the same procedure to recode
the other 3 variables.
REPEAT for all
3. Compute Cronbach’s Alpha
Select Analyze, Scale, and Reliability Analysis… to open the Reliability Analysis dialogue
box.
Select the 8 required variables to the Items: box. Ensure that Alpha is displayed in the Model:
box. Now click on the Statistics… button to open the Reliability Analysis:Statistics dialogue
box.
In the Descriptives for box, select the Scale and Scale if item deleted check boxes. In the
Inter-Item box, select the Correlations check box.

Then click Continue and OK to perform the analysis.

How did the scale perform? In the Corrected Item-Total Correlation column, which item has
the lowest value? opt1.1 and opt7.1 have the lowest values. Their Alpha if Item Deleted value
indicates that the alpha would change little if these two items were dropped to increase the
reliability of the scale slightly.

Reference
Coakes and Steed (2011), Chapter 15
 SPSS tutorial exercise 10 – STEP-BY-STEP

WORKING WITH SPSS OUTPUT

Objective: You have seen throughout these exercises that SPSS not only performs all of the
calculations for you, but it also provides a generally neat output of the results as well. These
output files can be saved or printed off and could be used in appendices to your marketing
research report. Of course, many of the output results include technical information that will
probably not be necessary to have reported in full. So you should decide which pieces of
information are the most useful to highlight in your report. It is often better to select just the most
important and relevant items to summarize in tables and figures. In this exercise we are going to
learn some basics of working with SPSS output and transferring this output into a Word
document.

Do the following:
1. Open the database MREHobbitData.sav
Click File, Open, and Data. Next to the Look in: box, click on the down arrow button to find and
select the MREHobbitData.sav file. Double-click on that file to open the data.
2. Obtain an output table
Select the Analyze menu. Click on Descriptive Statistics and then Frequencies… to open the
Frequencies dialog box. Select likely and click the right-arrow button to move this variable into
the Variable(s): box. Click OK to obtain the output. Examine this output.
3. Manipulate this output
Click on the Statistics branch of the output tree in the left pane. Now select View, Hide to
remove that table from the output. To return it again, click View, Show. To change the order in
which elements of the output appear, try clicking and dragging an item to a different location. To
insert a new title or text into the output, click on an object in the output which appears
immediately above your target location. Now click Insert, and either New Title, New Text or
one of the other options and you will see that a new area in the output has opened for you to
insert the text you wish to add. To extract an object, such as a table, to insert in a Word
document, click on the object you require and you will notice that a red arrows indicates your
selection. Now click Edit, Copy and then open a word document so that you can Paste the
object into the desired section of your report.
The following is pasted with ‘Keep Source Formatting’

How likely would it be for you to patronize this restaurant (new upscale restaurant)?

Cumulative
Frequency Percent Valid Percent Percent
Valid Very Unlikely 52 13.0 13.0 13.0
Somewhat Unlikely 77 19.3 19.3 32.3
Neither Likely Nor Unlikely 161 40.3 40.3 72.5
Somewhat Likely 38 9.5 9.5 82.0
Very Likely 72 18.0 18.0 100.0
Total 400 100.0 100.0

The following is pasted with ‘Merge Formatting’

How likely would it be for you to patronize this restaurant (new upscale restaurant)?

Cumulative
Frequency Percent Valid Percent Percent

Valid Very Unlikely 52 13.0 13.0 13.0

Somewhat Unlikely 77 19.3 19.3 32.3

Neither Likely Nor Unlikely 161 40.3 40.3 72.5

Somewhat Likely 38 9.5 9.5 82.0

Very Likely 72 18.0 18.0 100.0

Total 400 100.0 100.0

The following is pasted with ‘Keep Text Only’

How likely would it be for you to patronize this restaurant (new upscale restaurant)?

Frequency Percent Valid Percent Cumulative Percent

Valid Very Unlikely 52 13.0 13.0 13.0

Somewhat Unlikely 77 19.3 19.3 32.3

Neither Likely Nor Unlikely 161 40.3 40.3 72.5

Somewhat Likely 38 9.5 9.5 82.0

Very Likely 72 18.0 18.0 100.0

Total 400 100.0 100.0

4. Obtain an output chart and manipulate
Click Graphs, and Pie to open the Pie Charts box. Select the radio button for Summaries of
groups of cases and click Define to open the Define Pie: Summaries for Groups of Cases
box. We are going to plot a pie chart for the likely variable so select the likely variable to the
Define Slices by: box. In the Slices Represent box, click the N of Cases radio button. Click
OK to produce the graph and observe the results. Notice that, in the same way as in 3 above,
the same types of manipulations can be performed. Copy and Paste the pie chart to your Word
document.
In a word document… Right Click where you want to paste the chart…
The following Chart can then be moved, resized, etc