PU1 INTRODUCTION – WORKING WITH SPSS
IMPORTING DATA FILES
To perform statistical analyses with SPSS, data have to be entered into an SPSS data file, there are several ways
of doing this:
- Type in the data yourself;
- Import a file with data (File Import Data Excel)
SPSS will ask you whether it should read the variable names from the first row yes! Handy option.
(File Read Text Data select file type: Excel OK)
DATA FILE
Columns
= The vertical lines of numbers in a data file, which represent the variables.
Rows
= The horizontal lines in a data file, which represent the cases (participants).
MENU BAR
Name Applications
File Open and save a file
Edit Remove parts of the screen, re-order rows and columns, etc.
View Modify the appearance of data. For example, you can choose to view the ‘value labels’
instead of the raw scores.
Data Modify the data. For example, you can select certain cases, aggregate data from two
different files etc.
Transform Transform values. You can compute new values or recode existing values.
Analyse Offers you a range of different statistical analyses for descriptive and inferential statistics.
Graphs Offers you a range of different options to graphically display your data.
Utilities Inspect variables.
Extensions Provides extensions to the program.
Window Change between the data file and the output file.
Help In case you are lost.
Data view
= Data is displayed.
Variable view
= The program’s codebook. It shows what is behind the variables and scores displayed in the data view, and
allows you to specify many aspects of your variables and scores.
Adding scores to values
In the variable view you can add scores to the values (for example man = 0 and woman = 1). Go to variable and
click on the cell below ‘values’. Now specify the values as shown; value is specified as 0 (the score that you have
chosen to represent males), and Label as ‘man’. Click on add and repeat for women. Click on OK and check the
values in the variable view.
RUNNING A SPSS SYNTAX SCRIPT
Sometimes you have to run a syntax script which makes some important changes in the data file. You can do
this by opening the syntax file (in SPSS), select everything and press the green play button.
,PU2 DESCRIPTIVE STATISTICS
The main function of descriptive statistics is data reduction. Descriptive statistics like the mean, or standard
deviation, allow you to summarize or characterize a large number of numerical observations.
CONDUCTING DESCRIPTIVE STATISTICS
Click on Analyse in the menu bar (on top in the data editor). In the drop-down menu, select the function
Descriptive Statistics. This opens a sub menu withs several options: Frequencies, Descriptives, Explore,
Crosstabs, Ratio, P-P Plots and Q-Q Plots.
Frequencies
You use this information if you’re interested in the frequencies of one or several variables. This function is
commonly used for categorical variables (nominal measures), but in some cases for scale (continuous) variables
as well. Sometimes it may be useful to ask SPSS to produce a so-called histogram and a normal curve of the
frequencies of a scale variable. You can do this by clicking on Charts and tick the box Show normal curve on
histogram.
Descriptives
You use this information to calculate means, standard deviations, ranges, etc. This function can be used only for
scale variables. Click on Options, this opens a sub dialog box that gives you a range of options to perform
descriptive analyses. The options Mean, Std. deviation, Minimum, Maximum are select by default. Click on OK.
Frequencies again (Crosstabs)
You use this information if you’re interested in the combination of different variables. You can do this using the
option Crosstabs. Enter the variables of interest into the Row(s) and Column(s) boxes (doesn’t matter which
one you put where). Click on OK.
Split file
You use this information if you’re interest in the combination of different variables, but only one part of it. For
example, you are only interest in girls, therefore you split the file by the variable gender. Here is how you do
that (Data Split file Compare groups place the variable gender in the window on to the right click on
OK check data file bottom-right “split by gender”). IMPORTANT: The split file commando remains active until
you undo it. Thus, always remember to undo the selection by selecting ‘analyse all cases’ in the Split file menu
and hit OK.
REPORTING
Example:
In total, ... respondents from ... different countries completed the survey (.... males and ... females). Their mean
age was ... years (SD= ... ). Of these respondents, ... did a language related bachelor, ... did a media related
bachelor, and ... did a different bachelor (or did not do a bachelor yet). On average, respondents drink ... units
of alcohol per week (SD= .... ) and their average answer to the question whether they are good at math is ... (SD
= ... ). ... respondents do not drink any alcohol and half of the respondents do not drink more than ... units of
alcohol per week. Male/female respondents drink more units of alcohol per week (M= ... , SD=.... ) than
male/female respondents(M= .... , SD= ... ). Their mean score on the LexTale test of English was ... (SD= ... ).
, PU3 RECODING AND COMPUTING VARIABLES (SOME) EXPLORING DATA
There are several techniques by which you can recode the scores of a variable or use the scores to calculate
(‘compute’) new variables. You can find the functions for this in SPSS under the Transform menu: (Transform
Compute Variable OR Transform Recode into Same Variables/ Recode into Different Variables).
RECODE INTO SAME VARIABLES
There are many situations in which you want to recode your data. For example questionnaires often contain
questions that use 5- or 7-point scales which sometimes need to be reverse coded, because they are the
opposite (e.g. 1 = good, 5= bad and 1 = bad, 5 = good).
This option substitutes your original variables by the new, recoded variable. This means the original variable is
overwritten and therefor ‘gone’, so… DO NOT MAKE MISTAKES. Steps to take:
- Transform Recode into Same Variables select the variable you want to recode and drag it into the
Numeric Variables-screen.
- Click on Old and New Values to get access in the Recode into Same Variables: Old and New Values
submenu.
- To reverse the variable’s scores all 1’s must be recoded into 5’s and all 2’s into 4’s etc. Type 1 in the ‘Old
Value’ window and 5 in the ‘New Value’ window Click on Add.
- Repeat this for all scores (don’t forget to click Add after every recoding!)
- Click on ‘Continue’ and then ‘OK’. The variable in question is now recoded into the same one.
RECODE INTO DIFFERENT VARIABLES
The Recode into Different Variables- option works almost similar to the other Recode-option, except that
instead of overwriting the old variable, the new, recoded variable is added to the list of variables (at the end).
Steps to take:
- Transform Recode into Different Variables select the variable you want to recode and drag it into the
Numeric Variables-screen give the Output Variable a name [OriginalName]_r (r = recoded).
- Click on Old and New Values to get access in the Recode into Different Variables: Old and New Values
submenu.
- To reverse the variable’s scores all 1’s must be recoded into 5’s and all 2’s into 4’s etc. Type 1 in the ‘Old
Value’ window and 5 in the ‘New Value’ window Click on Add.
- Repeat this for all scores (don’t forget to click Add after every recoding!)
- Click on ‘Continue’ and then ‘OK’. The variable in question is now recoded into a different one.
COMPUTE VARIABLES
The Compute Variable function allows you to compute variables based on other variables. You can use this
function for example to calculate the mean of several variables that together form a scale, the sum of several
variables or the product of two variables.
Calculating the mean of two or more variables
A lot of calculating is possible with this function, but here we focus on its most common use: calculating the
mean of two or more variables. There are two ways to calculate the mean score using the compute function:
- By hand
(Open the Compute menu Give the computed variable a new name in the Target Variable box (top-left)
Drag the relevant variables into the Numeric Expression box. Between each of the variables, put a + sign