Quantitative Research Methods

Using SPSS 1

SPSS 25 is available in all labs. To enter the University networked version from the Windows desktop,
go to Start/All Programs/ IBM SPSS Statistics/IBM Statistics 25 and simply double-click.

You can obtain a student version of SPSS to download on your own laptop; details here:
https://stir.unidesk.ac.uk/tas/public/ssp/content/detail/service?
unid=83f99414f20148e987fa911d86ec3785&from=bc04099c-f613-4563-8f01-e6e589ca9833 (choose
“IT” and then “Software/ Apps for your own laptop”)

Entering data

In order to start familiarizing yourself with the SPSS software, you can start by entering data. The first
screen asks you whether you want to create a new dataset, open a sample file or a recent file, or run
the tutorial and get support. Since, at this stage, you do not have any existing data, select New
Dataset (top left corner) and then Open. The tutorials are also very helpful for understanding the
SPSS various functions.

You will now obtain the Data Editor window (Figure 1). This is for entering new data. The window is a
grid for creating a data matrix whose rows represent survey respondents (called ‘cases’ in SPSS) and
whose columns will contain the values of the variables for each case.

Figure 1 The Data Editor screen. Data View

Before entering any data, it is advisable first to define all the variables. In the Data Editor window click
on the Variable View tab on the bottom left hand side of the screen. This will give you the Variable
View screen (Figure 2).

1
 Figure 2 The Variable View

Table 1 suggests six variables relating to six questions addressed to a sample of respondents. Each
variable is given a short name and a more extended description may be given as a variable label.
Where variables are sets of categories, each category may be given a value label and a code. The
first variable is named Gender so type in Gender under Name in the first row of the first column.
These names must not exceed eight characters (unless you change the default of 8 in the Columns
column), they must begin with a letter, they must not end with a full stop and there must be no spaces.
Some words like and, not, by and all are reserved by SPSS and cannot be used.

Table 1

Variable name Variable label Value labels Code

Gender Sex of respondent Male 1
Female 2
Age Age in years

Ban Attitude to banning Strongly agree 5

advertising Agree 4
Neutral 3
Disagree 2
Strongly disagree 1
Sport Participation in sport Participates 1
Does not participate 2
Health Attitude to health Avoids fat 1
Does not avoid fat 2
Smoke Smoking behaviour Smokes 1
Does not smoke 2

As soon as you hit Enter or down arrow or right arrow, the remaining boxes will be filled with default
settings, except for Label. You can leave these settings as they are, but it is better to enter variable
labels in the Label column, so type in Sex of respondent in the Label column in the first row. The
restrictions on variable names do not apply to labels, so you can make them longer, use spaces and
so on.

2
If the variable is categorical (binary, nominal or ordinal), it is a good idea to enter value labels. Click in
the cell under Values against the variable Gender in the first row and click again on the little blue box
to the right of the cell. This will produce the Value Labels dialog box. See Figure 3.

Figure 3 Completed Variable View

In the Value box, type in 1 (the lowest code number) and in Value Label type in Male. N.B. To move
between boxes, click on the mouse in the appropriate box (or use the Tab key to move to the box
below). Click on the Add button and 1=“Male” will appear in the lowest box. Now add 2=Female using
the same procedure and then click on OK. Now name the other five variables, adding variable labels
and value labels. Note, however, that age is a metric variable and does not require value labels.

The default under Decimals is usually 2. If all the variables are integers, then it is worthwhile changing
this to zero. Simply click on the cell and use the little down arrow to reduce to zero.

Under Measure, you can put in the correct level of measurement – nominal (includes binary), ordinal
(ordered category or ranked) or scale (metric).

The completed Variable View is shown in Figure 3.

You are now ready to enter the data. Switch to the Data View and enter the data from Table 2. This
will normally be done row by row, i.e. questionnaire by questionnaire for survey analysis. Put the cell
highlight on the cell into which you wish to enter a value (begin top left) and simply type the number
(always enter the codes, not the value labels!). Move the highlight using the direction keys. Notice that
the process of entering is completed simply by moving the highlight to another cell. You could press
the Enter key instead.

3
Table 2

Gender Age Ban Sport Health Smoke

1 1 23 4 1 2 1
2 2 27 2 1 2 2
3 2 33 1 2 1 1
4 1 56 5 2 2 2
5 2 46 2 1 1 1
6 2 22 1 2 1 2
7 1 48 3 2 1 2
8 1 34 4 2 1 2
9 2 37 3 2 2 1
10 2 34 4 1 2 1
11 1 25 4 1 1 2
12 1 34 1 1 2 1
13 2 23 5 2 1 2
14 1 22 3 2 1 2
15 1 25 4 1 1 1

Notice that just above the grid is a white bar - this is the Data Editor. This shows your entry as you
type it and indicates the cell position on the left. The usual Windows editing functions are available,
for example you can cut, copy and paste in the usual way. To change a value in a cell once it has
been entered, simply highlight the cell, type in the new value and press Enter.

Note that SPSS assumes that all data matrices are rectangular. If you press Enter before you get to
the end of the second or subsequent rows a full stop is entered in each of the remaining cells in that
row. There can be no empty cells. If no value has been entered, the system supplies the system-
missing value, which is indicated by a full stop.

 Hint on future datasets

If, in future, all the values to be entered are integers (whole numbers), it is better to change the
default, which specifies the number of decimal places shown - this is normally set at two decimal
places. However, you must do this before you enter any data and before you put in variable labels.
From the menu bar at the top of the Data Editor window select Edit, then Options, then the Data tab.
Change Decimal Places to zero. Click on OK.

Saving your work (i)

Remember that SPSS does not have an automatic timed backup facility. You need to save your work
regularly as you go along. Use the File\Save sequence as usual for Windows applications. The first
time you go to save you will be given the Save As dialog box. Choose the folder in which you wish to
save your work. File\Exit will get you out of SPSS and back to the Windows desktop.

Data analysis

SPSS has some very powerful techniques at its disposal. However, for present purposes only the
very basic procedures will be explained. When analysing any dataset, you will need to begin by
obtaining one-way descriptive summaries for each of your variables, i.e. one at a time. Which
summaries are appropriate depends on the type of data. For non-metric (categorical) variables
(sometimes called ‘nominal’ or ‘ordinal’ scales) you need the Frequencies procedure. This is in the
Analyze/Descriptive Statistics drop-down menu from the menu bar at the top. So, click on Analyze,
then Descriptive Statistics, then Frequencies. The Frequencies dialog box will appear (see Figure 4).

4
 Figure 4 The Frequencies dialog box

All variables are listed in the left box. To obtain a frequency count for any variable simply transfer it to
the Variables box by highlighting it, then clicking on the direction button in the middle (or simply drag
across with the mouse). You can bring across as many variables as you like. Click on OK and you
obtain a frequency count for each variable along with Percent, Valid Percent and Cumulative Percent,
as illustrated in the Table 3 below for the variable Ban. Valid percent excludes missing cases if there
are any.

Table 3 Attitude to banning of advertising

Cumulative
Frequency Percent Valid Percent Percent
Valid Strongly disagree 3 20.0 20.0 20.0
Disagree 2 13.3 13.3 33.3
Neutral 3 20.0 20.0 53.3
Agree 5 33.3 33.3 86.7
Strongly Agree 2 13.3 13.3 100.0
Total 15 100.0 100.0

Where there are no missing cases the Percent and Valid Percent are the same. You can edit the table
to remove these columns if you wish. Just double-click on the table. This puts you into editing mode
and gives you a Formatting Toolbar. Highlight the figures in the column (drag the pointer down the
figures) and hit Delete. The column will disappear, and the table will close up. You can do the same
with the Cumulative Percent column. To get out of Edit mode just left-click outside the table area. With
the table highlighted (single click on the table - there will be a frame around it and a red arrow to the
left) you can select Edit and Copy and then Paste it into any other application like Word or
PowerPoint.

If you click on Charts in the Frequencies dialog box you obtain the Frequencies: Charts dialog box.
Simply click on Bar Chart and indicate whether you want the axis label to display frequencies or
percentage, click on Continue and then OK. This will give you a basic default bar chart in addition to
the frequencies table. To obtain other kinds of bar chart like ‘stacked’ or ‘clustered’ bar charts or

5
three-dimensional charts, you will need to select the Graphs drop-down menu. You can then choose
between a Chart Builder, which gives you a kind of chart wizard, or the dialog boxes that were on
previous versions of SPSS and which are now called Legacy Dialogs to give you the Bar Charts
dialog box. If you choose Chart Builder, click on OK. Drag the kind of chart you want into the Chart
preview box, then drag the variables across to the appropriate axes.

Once you have obtained your chart you can edit it by double-clicking in the chart area. This will give
you the SPSS Chart Editor. You can change the colours and a number of other chart features from
the editor. Close the Editor when you have finished. If you single-click on the chart area you highlight
it. If you now select Edit/Copy you can copy the chart into other applications.

For metric variables you can also use the Frequencies procedure, but in addition you can Click on the
Statistics button in the Frequencies dialog box to obtain a range of statistics. Just click the boxes as
required. Analyze/Descriptive Statistics/ Descriptives provides a quick way of obtaining a range of
common descriptive statistics for metric data, both of central tendency and dispersion - try it on age.
Put age into the Variables box and click on OK.

Saving your work (ii)

You can save the results of your analyses as presented in the ‘Output’ window but this is done
separately on a separate file than the data set one.

Using an existing data set

This exercise is based on the survey of table tennis players in Northern Ireland. Read Box 11.2 in
Kent (2007) for the background to the survey. You will need a copy of the questionnaire to do some of
the exercises below. The questionnaire is reproduced as Appendix 1 at the back of the book and is
also available in the Teaching material of this week on Canvas. The data have been entered into an
SPSS file Tabten2.sav.

To access these data, open the data file Tabten2 that you will find in the Teaching material of this
week on canvas.

Frequencies

When you get access to a dataset for the first time, the first thing to do is use the Frequencies
procedure to look at all the variables, the categories used for any non-metric variables, and the
distributions of the data in each variable. Tables rather than graphs are probably best for this purpose.
So, get a table for each of the 30 variables in the table tennis dataset. Notice, for example, that only 5
table tennis players play in the fourth division, so these could be combined into the previous category.
How to do this in SPSS is explained later under the Recode procedure.

Notice that in answer to the question ‘Who encouraged you?’, there are 35 ‘Missing’ and 85 ‘Valid’
cases in this table (Table 4). That means that 35 people did not answer this question. So, the 33 valid
cases who said they were encouraged by a friend represent 27.5 percent of the total sample of 120,
but 38.8 per cent of the 85 valid cases – this is the ‘Valid’ percent. The Cumulative Percent
accumulates the Valid Percents so that, for example, a total of 74.1 per cent were encouraged either
by a friend of by a parent. Since the scale here is nominal, the order is not important and reflects the
order in which the value labels were entered.

6
Table 4 Who encouraged table tennis players
Who encouraged you?

Cumulative
Frequency Percent Valid Percent Percent

Valid friend 33 27,5 38,8 38,8

Parent 30 25,0 35,3 74,1

Other relative 4 3,3 4,7 78,8

Teacher 2 1,7 2,4 81,2

Club leader 15 12,5 17,6 98,8

Other 1 ,8 1,2 100,0

Total 85 70,8 100,0

Missing System 35 29,2
Total 120 100,0

Crosstabs

The next stage in any data analysis is to look at the relationships between variables. For example, is
there any relationship between whether or not anybody else in the household plays table tennis (else)
and how often they play per week (play)? For this you need the Crosstabs procedure. This generates
crosstabulations for non-metric variables. Select Analyze/Descriptive Statistics/Crosstabs to obtain
the Crosstabs dialog box (it will be helpful to you maximize the dialog box to see all the details of the
variable names). Enter your dependent variable (play) in the Row(s) box so it will appear at the side,
and the independent variable else in the Column(s) box. Click on OK. Table 5 shows the output.

Table 5 How many times do you play per week? * Does anybody else in
your household play T.T? Crosstabulation
Count

Does anybody else in

your household play
T.T?

yes no Total

How many times do you play once 15 28 43

per week? twice 22 30 52

three times 10 8 18

four or more 5 2 7
times
Total 52 68 120

This table suggests that there may be a tendency for those who have nobody else in the households
who plays may play less often. But how strong is this tendency? If the independent variable is at the
top, the column percentages may help. To obtain column percentages, click on the Cells button in the
Crosstabs dialog box to obtain Crosstabs: Cell Display. Click on Column in the Percentages check
box, then on Continue, then on OK. Notice that the frequencies (called ‘Count’) and the percentages

7
to one decimal place are shown in each cell. The frequencies can be edited out if you want to display
just percentages. Notice that a higher percentage of those who have somebody else in the household
who plays table tennis play three or more times a week compared with those where there is nobody
else (see Table 6). Conversely, a higher percentage of those with nobody else who plays, play once
per week, but the difference for twice a week is very small.

Table 6 How many times do you play per week? * Does anybody else in your
household play T.T? Crosstabulation
% within Does anybody else in your household play T.T?

Does anybody else in your

household play T.T?

yes no Total

How many times do you play once 28,8% 41,2% 35,8%

per week? twice 42,3% 44,1% 43,3%

three times 19,2% 11,8% 15,0%

four or more 9,6% 2,9% 5,8%

times
Total 100,0% 100,0% 100,0%

If you suspect that the relationship may differ for males and females, you can obtain a 3-way table.
Put a variable that you wish to use as a ‘control’, e.g. gender, into the Layer box in the Crosstabs
dialog box and click on OK. You will, in effect, obtain two crosstabulations, one underneath the other,
one of males and one for females. You can layer by more than one variable, but the frequencies in the
cells get very low and the whole thing becomes more difficult to interpret. Try it and see what
happens.

Measures of association

Looking at percentage differences can give quite a lot of information about the nature of the
relationships between two variables, but not an overall summary measure of the degree of association
between the two. SPSS offers no fewer than nine different coefficients of association, four for when
both variables are nominal, four for when they are ordered categories and one for a mixture of
categorical and metric. These measures are explained in Kent, 2007, pp. 345-360 and Kent, 2015, pp.
122-131.

The most useful and generally applicable statistic of these nine is Cramer’s V. It can be used on any
size of table and for binary, nominal or ordered category variables or any mixture of these. When
either (or both) is binary, then Cramer’s V is identical to a statistic called Phi and SPSS puts both of
these together. To obtain it, in the Crosstabs dialog box, click on Statistics to get the Crosstabs:
Statistics dialog box and select Phi and Cramer’s V. Click on Continue then OK and you should obtain
Table 7.

Table 7. Phi and Cramer’s V for Table 6

Symmetric Measures

Value Approx. Sig.

Nominal by Nominal Phi ,196 ,202

Cramer's V ,196 ,202

N of Valid Cases 120

8
Cramer’s V is identical to Phi (else is binary) at 0.196, which is a very low degree of association.
‘Symmetric’, in SPSS language, means that the measure is not directional so will give the same value
if the variables were entered into the Crosstabs with play at the top. ‘Approx.Sig’ is what is called the
p-value. If the 120 cases are a random sample and we can safely ignore all sources of error other
than random error then there is a 20.2 per cent chance of getting a value of Cramer’s V as a result of
random sampling error. In marketing and sociology, the critical p-value is usually taken to be 0.05. If
the result is less than 0.05, we can consider that the achieved coefficient of association is unlikely to
be a consequence of random sampling fluctuation since there are fewer than 5 chances in 100 of this
happening. The result is then said to be ‘statistically significant’ – there is an association that is
probably not random sampling error. For Table 7, p. is 0.202, which is much larger than 0.05, so, it is
not statistically significant – the resulting Cramer’s V could easily have arisen by chance.

9
 Using SPSS 2

For this exercise, you will need the same table tennis data set we used in Exercise 1 (i.e. Tabten2).

Recode

Where there are several or indeed many values in a nominal or ordered category variable or where
the set of cases in a dataset is small, some of the frequencies in some of the categories will often be
small. If the variable is ordered category, it may make sense to add together the frequencies of
adjacent categories. If the variable is nominal, the groupings need to make sense in some way. In Q8
of the questionnaire provided, for example, respondents are asked to rate the importance they attach
to various aspects of playing table tennis. For social benefits, the responses are as in Table 1.

Table 1 Social benefits of playing table tennis

Frequency Percent

Valid unimportant 11 9,2

fairly unimportant 18 15,0

neither 41 34,2

fairly important 37 30,8

very important 13 10,8

Total 120 100,0

There are relatively few who have indicated ‘very unimportant’ or ‘Very important’. It might be helpful
to add together ‘unimportant’ and ‘fairly unimportant’ into a new category, and to add together ‘fairly
important’ and ‘very important’ into another category.

To do this select Transform/ Recode into Different Variables. Move ‘Social benefits’ into the Input
Variable -> Output Variable box. Now click on Old and New Values. We need codes 1 and 2 to
become code 1 so in the Old Value dialog area on the left click on the first Range radio button and
enter 1 through 2. In the New Value dialog area on the right enter 1 in the Value box and click on Add.
This instruction will now be entered into the Old -->New box. Code 3 we want to change to 2 so click
on the Value radio button under Old Value and enter 3. Now enter 2 under New Value and click on
Add. We now want codes 4 and 5 to be code 3. Click on the Range radio button and enter 4 through
5. Under New Value enter 3 and click on Add. Click on Continue. Give the Output Variable a name in
the Name box, for example, socben3 and click on Change then OK. The new variable will appear as
the last column.

You can now do this with the other four variables in Q8. You will not need to enter the instructions
again - they remain until you use Remove to delete them!

To add value labels for the new variable, change to the Variable View. Click in the right corner of the
Values cell in the appropriate row and obtain the Value Labels dialog box. Enter 1 in Value and
Unimportant under Value Label and click on Add. Now enter 2 in Value and Neither under Value
Label and click on Add. Finally, enter 3 in Value and Important under Value Label and click on Add.
Now click on Continue and OK. You can now check this out using Analyze/Descriptive Statistics /
Frequencies procedure.

Hint. It makes sense for the most positive category to be allocated the highest code value, so then
code of 3 goes to 'Important'. Unfortunately, SPSS will, in table outputs, put the lowest code (and the
least positive category) at the top as in Table 1. However, if you click on Format in the Frequencies

10
dialog box, you can select Descending values and the order will be reversed with 'Important' at the
top.

Compute

Where respondents have been asked to rate several aspects of a dimension and the ratings may
reasonably be considered to be roughly equally spaced, then the researcher may wish to create a
total score for that set of items for each respondent. Q9, for example, asks respondents how satisfied
they are with various elements of table tennis like practice facilities, competition facilities and so on. If
the codes 1-5 may reasonably be considered to be ratings (and watch that the most positive category
has been given the highest code), then a total overall satisfaction score for each respondent may be
created. To do this, select Transform/ Compute Variable, then put the 5 items practice, compfac,
admin, coach and compete into the Numeric Expression box with a + between them. Don't forget to
give the Target Variable a name like totalsat. Click on OK and a new variable will be created. Check
this out in your Data View.

The total scores will vary from 5 to 25. You can now use Recode to group the scores into a limited
number of categories, e.g. 5-12 = 1 (low), 13-15 = 2 (medium), and 16-25 = 3 (high). The new variable
can then be crosstabulated against any other categorical variables. Alternatively, in the Numeric
Expression dialog box (Compute variable window), you can put the 5 variables between brackets and
enter /5 to give an average score for each person.

Reliability Analysis

When Compute has been used on a multi-item scale, it is important that responses should be
consistent so that, for example, there is a tendency for those who are satisfied with table tennis in
N.Ireland will be satisfied with all the elements and vice versa. A measure of consistency that is
commonly used is Cronbach's coefficient alpha. This takes the average correlation between the items
and adjusts for the number of items. Reliable scales are ones with a high average correlation on a
large number of items. The coefficient varies from zero for no reliability (or consistency) up to 1 for
maximum reliability.

For Q9 select Analyze/Scale/Reliability Analysis. Check that Model is set at Alpha. Select the 5 items
in Q9 and put them into the Items box. Click on OK. Note the value for Alpha. Most researchers
suggest that values of alpha over 0.7 mean that it makes sense to add up the items. The result for Q9
is 0.73, so it looks as though it should be acceptable.

If under 0.7 then click on the Statistics button at the bottom of the Reliability Analysis box and select
Scale if item deleted, then Continue, then OK. You can now see whether any items are limiting the
overall score for alpha.

Multiple Response

There are often questions in a survey that allow respondents to select as many responses as apply to
them. For the purpose of analysis, each response is treated separate (binary - ticked or not ticked)
variable. These responses now need to be analysed together. Q19, for example asks respondents
what other sports they play. They may select none or several of these. Select Analyze/Multiple
Response/Define Variable Sets. Transfer foot, tennis, squash, badmin from the Set definition box to
the Variables in Set box. Click in the Dichotomies Counted value radio button in the Variables Are
Coded As box and enter the value to be counted as part of the set, e.g. if the value 1 has been
entered on the questionnaire to indicate that this item has been selected, then enter 1 in this box.
Don’t forget to give a name to the set. Entering a label is optional. Click on Add, then Close.

Several multiple response sets may be defined at the same time. These are then listed in the Mult
Response Sets box. Note that these do not appear as new variables. To access them you need to
reselect Analyze/Multiple Response and either Frequencies or Crosstabs. The sets so defined may

11
now be used either to show frequencies or they may be used in Crosstabs. You will for the latter,
however, need to define the value range of the other variable from the Define Range button.

The Result for Q19 is shown in Table 2

Table 2 $Sport Frequencies

Responses Percent of
N Percent Cases

$Sporta Play football 82 58,6% 73,9%

Play tennis 29 20,7% 26,1%

Play squash 17 12,1% 15,3%

Play badminton 12 8,6% 10,8%

Total 140 100,0% 126,1%

a. Dichotomy group tabulated at value 1.

The Case Summary in SPSS shows that there are 9 missing cases out of the 120. These are those
who, presumably, do not play any of the other sports. There are 140 responses in total, the largest
number indicating football, which accounts for 58.6 per cent of responses.

Correlation and regression

Where two or more variables are metric or may reasonably regarded as metric, then the pattern of
relationships between the variables may be summarised using the related ideas of correlation and
regression. This pattern is best seen in a scattergram. Two metric variables in the table tennis dataset
are the amount spent on table tennis in the last 6 months (Q18) and the age at which they began
playing table tennis (Q6). There is reasonable expectation that the later respondents began play the
more they may have spend on it recently. Select Graphs/Chart Builder/OK/Scatter/Dot. Select Simple
Scatter and drag to Chart Preview. Drag agebegan to the X-axis and spend to the Y-axis and click on
OK (it will be helpful to maximize the dialog box to see all the details of the variable names ). The
result is shown in Figure 1.

12
 Figure 1 A scattergram of agebegan and spend

You can see that there is a tendency for recent spend to be higher for those who began later in life
and to be lower for those who began at an early age. Correlation provides an index of the extent to
which two metric variables covary. It measures how near, on average each, each data point is to a
hypothetical line that runs through the points in a way that minimises the distances between each data
point and the line. To get SPSS to produce such line (called a regression line), go into Edit mode (by
double clicking on the graph) and click on the Add Fit Line at Total icon (there is only one icon with
this label). The standard measure of correlation is one called Pearson's r. to obtain it for these two
variables go to Analyze/ Correlate/ Bivariate. Transfer agebegan and spend into the Variables box,
check that Pearson is ticked under Correlation Coefficients and click on OK. Note that r = 0.808,
which is quite high. The p-value is 0.000 so the result is statistically significant.

Table 3 Correlations

Approximately
how much have
At what age did you spent on
you take up T.T.in the last 6
table tennis? months?

At what age did you take up Pearson Correlation 1 ,808**

table tennis? Sig. (2-tailed) ,000

N 120 120
Approximately how much Pearson Correlation ,808** 1
have you spent on T.T.in the Sig. (2-tailed) ,000
last 6 months? N 120 120

**. Correlation is significant at the 0.01 level (2-tailed).

13
Regression takes correlation a bit further, so the researcher can use the values on one variable to
predict the other, using the regression line. This is described by the general formula Y = a + bX. Put in
any age at which they began playing table tennis (the X value) and you can predict how much they
are likely to have spent on it in the last 6 months. SPSS has a separate regression analysis. Select
Analyze/Regression/Linear. Put agebegan into the Independent(s) box and spend into the Dependent
box and click on OK. The result is shown in Table 3.

Table 4 Coefficientsa

Unstandardized Standardized
Coefficients Coefficients

Model B Std. Error Beta t Sig.

1 (Constant) -2,691 3,928 -,685 ,495

At what age did you take up 4,253 ,286 ,808 14,871 ,000
table tennis?

a. Dependent Variable: Approximately how much have you spent on T.T.in the last 6 months?

This gives you the constant, -2.69 and the slope, 4.25, which go into the equation: Y= -2.69 + 4.25X

Exercises developed by Ray Kent

References

Kent, R. (2015) Analysis Quantitative Data. Variable-based and Case-based Approaches to Non-
Experimental Datasets. London: Sage Publications.

Kent, R. (2007). Marketing Research: Approaches, Methods and Applications in Europe, Thomson
Learning.

14