Using SPSS (IBM SPSS Statistics)
1. Introduction
5. User Library
1. Introduction
This Help Note describes how to use SPSS on the SSC Network.
Running SPSS
SPSS may be run from the SSC Network menu as follows:
Start > All Programs > All Programs > IBM SPSS Statistics > IBM SPSS Statistics 19
Students can also run SPSS remotely from MyVLab using their web browser at myvlab.uwo.ca.
Getting Help With SPSS
There are a number of resources from which you can use to get assistance using SPSS:
- SPSS has included an excellent tutorial which is available from the Help menu.
- A Statistics Coach is also available from the Help menu.
- Each dialog box has a help button which provides on-line context sensitive help.
- Reference the syntax examples on the web:
ssnds.uwo.ca/statsexamples/ - Western Libraries has SPSS software manuals located in the Map and Data Centre (Room 1051, SSC) that can be consulted within the centre.
Windows, Menus and Toolbars
When you start SPSS, the Data Editor window displays the contents of the working data file (see below). Other windows are opened when required.
Data Editor:
Display and edit data
Chart Editor:
Edit graphs
Viewer:
Display statistical results and charts
Syntax Editor:
Display and edit SPSS commands
Each window has its own menu bar and toolbar. To see a brief
description of each tool, pass the mouse over the relevant icon on the
toolbar. Multiple windows of a particular type (excluding the Data
Editor window) may be open concurrently. All activity for a particular
window type occurs in the default window. Click the
tool to designate another window as the default of that type.
Information area
The status bar, located at the bottom of each window, contains information pertinent to the current activities. These include the status of the SPSS processor, data filters and split files.
Options
There are a number of general options (via the Edit menu) which you may wish to modify. For example, to display SPSS commands in the log of the Viewer window select the Viewer tab and click the check box Display commands in the log.
2. Working with Data
Data can be entered directly into the Data Editor window or retrieved from a file. You can also perform a wide range of transformations on your data. Keep in mind that any changes you make to the data will not be available the next time you run SPSS unless you save your data.
Data Entry
The Data Editor provides two views of your data. The Variable View allows you to defines your variables. (ie., data type, variable label, value labels, missing values, column format). Data may be entered in the Data View.
SPSS Data Files
Data stored as an SPSS data file (*.SAV) may be opened directly into the Data Editor. This is the most common type of data file used with SPSS.
Other Types of Data Files
SPSS can read many types of data files including text files, spreadsheets and databases.
Transformations
You can create new data values based on numeric transformations of existing variables. Of the many transformations available, recoding values and numeric computations are the most common.
To change variable codes select Recode from the Transform menu. Typical uses for recoding might be to reverse code a 5-point Likert scale where 1 is recoded to 5, 2 is recoded to 4, 4 to 2, and 5 to 1; or to collapse a range of values into a single value to produce a grouping variable. By recoding into the same variable you will replace the original values with the recoded values. If you wish to keep the original values simply recode into a different variable.
New variables may be computed from old ones by selecting Compute from the Transform menu. You may use arithmetic expressions, relational operators, and various kinds of predefined functions in your computations.
Missing Data
Observations which are unknown or not specified are assigned the system-missing value. SPSS uses a period to indicate missing data for numeric variables and a blank for string variables. In the example above, the third case is missing a subject label and subject A37 is missing the second test score.
Sometimes it is useful to know why data is missing. You can define an observation as being missing via the Variable View within the Data Editor. These values are termed user-missing. By specifying subject H19's third test score as user-missing, the -9 is flagged for special treatment and is excluded from most calculations.
Care should be taken when working with missing data since missing data is propagated through arithmetic expressions. For example, if you calculated the average test score for subject A37 by adding the three test scores and then dividing by 3, SPSS would produce the system-missing value. All SPSS statistical functions include a strategy whereby only valid values are used in the computation. The MEAN function would produce the value 78, the average of the two non-missing test scores for subject A37.
Many of the statistical analyses provide alternative methods for dealing with missing data. Remember you can see a description of any element in a dialog box by right-clicking on it.
3. Case Selection
Transformations may be computed conditionally. For example, you may
wish to use a different transformation when computing a new variable,
depending on the value of an existing variable. To specify the
condition, click the
button in the Compute Variable dialog box.
A subset of cases may be selected for inclusion in an analysis. From the Data menu choose Select Cases and make your selection. SPSS will then create a filter variable, place a hash mark over the case numbers of those cases not included in the selection, and place the message "Filter On" in the status bar. To return the focus to the entire data set select All cases in the Select Cases dialog box.
To carry out a number of analyses on different groups of data, select Split File from the Data menu and choose a variable on which to base your groups. SPSS will sort your data by the grouping variable unless you tell it that your file is already sorted. The message "Split File On" will appear in the status bar. Select Analyze all cases, do not create groups to return the focus to the entire data set.
4. Using the Syntax Editor
Every time you click the
button in a dialog box you are submitting commands (or syntax) to the
SPSS processor. Below each
button is a
button. If you click the button
the commands are pasted into a Syntax window. At this point you can edit
the commands in the syntax editor, save them, or run the commands. These
operations are discussed below.
Editing Commands
When editing command syntax you must follow a few simple rules. Each
command should begin on a new line and end with a period. Separate
subcommands with a forward slash (the slash before the first subcommand
is optional). Use the Help menu if you wish to read the complete set of
syntax rules. To see the syntax chart for any command click the
button.
Saving Commands to a File
Command syntax may be saved to a file for future use (select Save from the File menu). The file type .SPS indicates an SPSS syntax file.
Running Commands
To run all of your commands select All from the Run menu. If you wish to submit only a portion of your syntax to the SPSS processor choose Selection from the Run menu. Note that transformations are left in a pending state until SPSS encounters an executable command (such as FREQUENCIES or T-TEST) or the EXECUTE command.
In the above example a new variable is created for the second test score using mean substitution. First, the compute command assigns the values of test2 to newtest2. Then a (conditional) command replaces missing values in test2 with the mean of non-missing test scores. Finally, the execute command is used to run the transformations.
5. User Library
Two libraries, Course Library and Sample Library are located on the SSC Network via the L: drive. The Course Library is a collection of folders and files organized by department and course offering. Instructors may use this service to provide students with easy access to on-line files such as SPSS data files, syntax files or course notes.
Instructors using the ITS Genlabs should post their files on their course WebCT area.
Also from this web page:
Related
Contact
.: Heather Stevens
Network & Data Services
519 661-2111 ext. 83595
hstevens@uwo.ca
