Victoria University Business School
SPSS Instructions
BEO6000 Data Analysis for Business
Prepared by
Dr Michelle Fong
SPSS Spreadsheet
The rows of the data file are the information of each of the subjects (such as a company or respondent). These are called cases. The columns are the variables of interest.
Data Files
Usually the most convenient way to construct a data file is in a spreadsheet such as Excel or Lotus. SPSS will read the data directly from a spreadsheet, however it will also read a text file written in a word processor. If you are using a spreadsheet as recommended, place the name of each variable at the top of the columns. Keep the names simple (for example Q1, Q2, … or VAR1, VAR2, …
To make SPSS read the data file,
click on file
click on open
click on data
click on Excel
click on look in to indicate the drive containing the data file
click the file name
click on open
click on read variable names
click OK
SPSS will then read your Excel data file.
As alternatives, you may enter the data straight into the SPSS data spreadsheet or paste it in from Excel.
Formatting the Data
The variable names used in the data file are often too cryptic. A person reading SPSS output would have no easy way of knowing what Q2 is unless they had the questionnaire at hand. To clarify, we can attach variable labels. Even when we have attached variable labels, person reading SPSS output would have no easy way of knowing what the codes 1 and 2 meant for Q2. To clarify, we attach value labels to show that 1 means “male” and 2 means “female”.
For SPSS version 26,
click on view
click on variables
write variable name in labels cell (eg. sex of respondent)
click on values cell
click on grey box at the right hand end of the cell
type first code in value (eg. 1)
type label in value label (eg. male)
click add
continue until all value labels have been added
click continue
click OK
Saving the Formatted Data
Obviously it is tedious to do this every time we want to start up SPSS to do another task with the same data. We can save the data together with the value labels and variable labels for later use. To do this,
click file
click save as
click save in to indicate which drive you wish to save to
in file name type the name you wish the file to have
click OK
This will save the data and whatever formats you have attached to it as a SPSS file with a SAV extension. Next time you want to use the data, just open the SAV file.
Re-coding Data
Sometimes we want to recode data. For example, we might want to use age groups such as 1=“under 30”, 2=“31 – 50” and 3=“51 and over”. To do this,
click the top of the column you wish to recode to highlight it
click on transform
click on recode
click on into different variables
click on the variable you wish to recode
click on right arrow
in name define the name of the new variable
in label attach a variable name (optional)
click old and new values
on the old values side click range
enter the boundaries for the first data range
on the new values side enter the code for the first range at value
click add
repeat (k) to (n) until all ranges have been entered
click continue
click change
click OK
SPSS will add a new column containing the new recoded variable. The next task is to attach value labels to the codes you have just created. This was explained earlier.
Selection of Data
Often we want to perform task on subsets of the data. For example we may wish to examine the responses of males over the age of 30. For this we use the select if command. To do this,
click on data
click on select cases
click on if condition is satisfied
click on if
click on the variable you wish to select (e.g. q1)
click on right arrow
type in condition you wish to be satisfied (e.g. <20)
click continue
click OK
When you a statistical procedure SPSS will only consider the cases that meet the condition you have specified. Note that once you have specified a select cases it will apply to all the tasks that follow. Obviously this could be a problem! To undo selected cases, repeat steps (a) and (b) and click on all cases and OK.
Creating New Variables
Sometimes we want to compute a new variable based on information already in the data file. For example, we may wish to use the year in which each respondent joined the company. Suppose we call the new variable START = 1998 – q4. To do this,
click on transform
click on compute
type the name of the new variable under target variable (e.g. START)
type the formula for this new variable under numeric expressions (e.g.1998–q4)
click OK
SPSS will add a new column containing the new variable.
Statistical Procedures
We now describe how SPSS is used to perform statistical procedures.
Summary Statistics
To perform a statistical task such as obtaining descriptive statistics, the following steps are used
click on analyse
click on descriptive statistics
click on descriptives
click on the variable that you wish to use
click on right arrow
click on OK
The SPSS output will provide such statistics as the mean value for the chosen variable.
Frequency Distribution
Another statistical task you may wish to perform is to obtain a frequency distribution. To do this,
click on analyse
click on descriptive statistics
click on frequencies
click on the variable that you wish to perform the frequency
click on right arrow
click on OK
The SPSS output will give a table of frequency of values for the particular variable used.
Percentile & Quartile
click on analyse
click on descriptive statistics
click on frequencies
click on the variable of interest
click on statistics
click on quartiles
click on percentiles (enter the percentile value into the box) and click add
click on continue
click on OK
SPSS does not provide Interquartile Range, you have to calculate it.
Obtaining z scores
click on analyse
click on descriptive statistics
click on descriptive
click on the variable of interest
click on save standardized values as variables
click on OK
A column (usually the last column) containing the z scores will appear in the SPSS spreadsheet.
Cross Tabulations
click on analyze
click on reports
click on crosstabs
click on the variable you wish to be the row variable (e.g.q2)
click on the right arrow for the row box
click on the variable you wish to be the column variable (e.g.q3)
click on the right arrow for the column box
click on the cell button at the bottom of the dialogue box
click on row, column and total in the percentage section
click OK
Case Summaries
click on analyse
click on descriptive statistics
click on case summaries
click on the variable that you wish to the variable (e.g. q3 for expenditure)
click on right arrow for variable box
click on the variable that you wish to the grouping variable(s) (e.g. q1 for sex)
click on right arrow for grouping variable(s) box
click on statistic to select the measure(s) that you want to use
click on option to enter the heading(s) for the output
ensure that display cases is not ticked (deactivate it)
click OK
Chi-square Goodness of Fit Test
click on analyze
click nonparametric tests
click legacy dialogue
click chi-square
select variable into the test variable list box
click OK
Chi-square Tests of Statistical Independence
click on analyze
click on descriptive statistics
(c) click on crosstabs
click on the variable you wish to be the row variable (e.g.q2)
click on the right arrow for the row box
click on the variable you wish to be the column variable (e.g.q3)
click on the right arrow for the column box
click on statistics
click on chi-square
click on continue
click OK
One Sample t-Test
click on analyze
click on compare means
click on one sample t-test
click on the variable you wish to use (e.g. operating revenue)
enter test value
click option
enter confidence interval percentage
click on continue
click OK
Note: The above SPSS instructions can be used to obtain confidence interval estimation by entering ‘0’ value into the test value field.
Z and t tests for Independent Samples
click on analyze
click on compare means
click on independent samples t-test
click on the variable you wish to use (e.g. q1 for age)
click on the right arrow for test variable box
click on the variable you wish to use for grouping variables (e.g. q2 for sex)
click on the right arrow for grouping variables box
click on define group
enter values for group1 and group2 (e.g. 1 for group1 and 2 for group2)
click option to check/enter confidence interval
click on continue
click OK
The Wilcoxon Rank Sum Test
This is a non-parametric test similar to a t test for two independent samples.
click on analyze
click on nonparametric tests
click on independent samples
click on objective
click on customize analysis
click on fields
click on the variable you wish to use (e.g. q1)
click on the right arrow for test fields box
click on the variable you wish to use for grouping variables (e.g. q2)
click on the right arrow for grouping variables box
click on define group
enter values for group1 and group2 (e.g. 1 for group1 and 2 for group2)
click on settings
click on choose tests
click on customize tests
click on Mann-Whitney U (2 samples)
click on test options
enter the significance level
enter the confidence level
click RUN
This procedure provide a Mann-Whitney U test as well as the similar Wilcoxon rank sum test.
Paired Sample t-Tests
Here we test the mean of the difference between two variables, for example Q7 and Q8.
click on analyze
click on compare means
click on paired-samples t-test
click on chosen variable for variable 1 (e.g. Q7)
click on chosen variable for variable 2 (e.g. Q8)
click on the right arrow
click OK
Wilcoxon Signed-Rank Tests for Paired Sample
This is a non-parametric test similar to a t test for paired samples.
click on analyze
click on nonparametric tests
click on 2 related sample
click on objective
click on customize analysis
click on fields
click on chosen variable for variable 1
click on chosen variable for variable 2
click on settings
click on choose tests
click on customize tests
click on Wilcoxon match-paired signed-rank (2 samples)
click on test options
enter the significance level
enter the confidence level
click RUN
One-Way ANOVA
Here we might compare the mean ages of people with different Likert scale scores for Q7.
click on analyze
click on compare means
click on one-way ANOVA
click on chosen variable for dependent list (e.g. Q1)
click on right arrow
click on chosen variable for factor (e.g. Q7)
click on right arrow
click on option
click on descriptive and homogeneity of variance test
click on contrasts
enter the respective coefficient for the first, second and third group accordingly.
click on add
click on post hoc to specify the type of confidence interval
click OK
Kruskal-Wallace Test
This is a non-parametric test similar to One-Way ANOVA.
click on analyze
click on nonparametric tests
click legacy dialogs
click on k independent samples
click on chosen variable for test variable list (e.g. Q1)
click on right arrow
click on chosen variable for grouping variable (e.g. Q7)
click on right arrow
click on define range
enter minimum and maximum values (e.g. 1 and 5)
click on continue
click on Kruskal-Wallace H in test type box
click OK
Two-Way ANOVA
click on analyze
click on general linear model
click on univariate
click model
select full factorial
click on chosen variable for dependent (e.g. Q1)
click on right arrow
click on chosen variable for factor(s) (e.g. Q2 and Q3)
click on right arrow
click plots for profile plot
click on chosen variable into horizontal axis (e.g. Q2)
click on chosen variable into separate lines (e.g. Q3)
click add
click continue
click on post hoc to specify the type of confidence interval
click on continue
click OK
Randomized Block Design
For a random block design, follow the steps for a two-way ANOVA and add the following.
click on model
click custom
change interaction to main effects
Correlation
click analyze
click correlate
click bivariate
click on chosen variables
click pearson or spearman in correlation coefficients
click two-tailed in test of significance
click flag significant correlations
click OK
Regression Analysis
To perform a linear multiple regression of the general form Q7 = f (Q1, Q2, Q3) the procedure is as follows.
click on analyze
click on regression
click on linear
click on chosen variable for dependent (e.g. Q7)
click on right arrow
click on first chosen independent variable for independent(s) (e.g. Q1)
click on right arrow
repeat steps (f) and (g) for other independent variables (e.g. Q2 and Q3)
click on statistics to select diagnostics suitable for the particular model
select estimates, model fit, collinearity diagnostic and descriptives
click on continue
click OK
Comparing least squares models (linear vs quadratic vs exponential models)
click on analyze
click on regression
click on curve estimation
click on chosen variable for dependent
click on right arrow
click on first chosen independent variable for independent(s)
click on right arrow
select models
click OK
Saving SPSS Output
To save SPSS output, firstly you must be in the SPSS ouput window. The following steps are then used to save the output.
click on file
click on save as
click save in to indicate which drive you wish to save to
in file name type the name you wish the file to have
click on save
This will save the output as a Navigator file (*.spv). This will not allow you to import your results into Word directly but you can retrieve them later in SPSS.
SPSS Output in Word
Importing the output file in HTML format into Word often produces a messy result. It is easy to copy each table that you really want to Word and edit it there. SPSS often produces more tables than you need, and in some that you do need it provides too much information. Copy only the tables you need and edit out superfluous material.
To copy a table to Word, click anywhere in the table in SPSS, click edit and copy, go to Word and paste. In Word, delete any unwanted material then anywhere in the table use table, table autoformat and simple 1. Then make the table conform with the margins of the document, space the columns, check the column justifications, use bold and/or bold italics for headings and sub-headings and use caption to attach a suitable table heading.
This is how it appears in SPSS (serviceable but no style).
And how it could look.
Table 1: Perceptions of Friendliness at the Workplace by Sex
|
|
|
Sex |
Total |
||
“Workforce is friendly” |
|
male |
female |
||
|
|
strongly agree |
4 |
– |
4 |
|
|
|
agree |
4 |
2 |
6 |
|
|
|
no view |
5 |
– |
5 |
|
|
|
disagree |
4 |
3 |
7 |
|
|
|
strongly disagree |
6 |
2 |
8 |
|
|
|
Total |
23 |
7 |
30 |
|
Special Note
Remember that, assuming the data has already been collected, there are three basic tasks:
Decide what statistical procedures have to be performed. What are the hypotheses and what sorts of tests are necessary.
Organize the data and use SPSS to perform the statistical procedures.
Interpret the output from SPSS. State the conclusions of the analysis. This must be concise and written in plain English.
- Assignment status: Already Solved By Our Experts
- (USA, AUS, UK & CA PhD. Writers)
- CLICK HERE TO GET A PROFESSIONAL WRITER TO WORK ON THIS PAPER AND OTHER SIMILAR PAPERS, GET A NON PLAGIARIZED PAPER FROM OUR EXPERTS
