Welcome to the world of SPSS (Statistical Package for Social Sciences). As you probably have already noticed, you will not be computing all statistics by hand for this course and the others you will be taking. If you have taken Statistics you are probably familiar with Minitab, another statistical package. If Minitab confused or frustrated you, you will be happy to hear that SPSS is a much easier to work with and has better graphics.

There are currently two versions of SPSS on the Acadia network: Version 6.1.4 and Version 8 (which is the version installed on the laptops). The instructions here are for Version 8, but many of the commands will also work for earlier versions.

This document will provide you with information pertaining to the following common statistical tests (many of which you will be using in this course):

There are many more statistics that SPSS can calculate, but they are beyond the scope of this document. Refer to the help dialogs in SPSS, or consult one of the manuals for assistance in conducting other statistics.

SPSS is a network application, so you must be on campus and logged into the Acadia network through an Ethernet connection to access it. If you are using Windows 95 on a laptop, click on start and move the pointer up to network applications. When that menu opens, move the pointer over to SPSS 8.0 to open SPSS. If you are in a computer lab, these same instructions may apply, or there may be a shortcut on the desktop. If you cannot find the program, consult the on duty lab consultant for assistance.

On successful opening, you will see the SPSS start up window shown in figure 1. This window disappears when SPSS is ready.

figure 1: SPSS startup window.

SPSS opens two windows. The first is the new data window, the second is the syntax window. The syntax window allows you to program SPSS to run its analyses rather than use the pull down menus. For more information on syntax windows, consult an SPSS manual.

SPSS has both a pull down menu and a graphical tool bar. Each of these runs your commands. Figure 2 shows you the pull down menu and the tool bar buttons, as well as the functions they serve.

figure 2: Pull-down menu and tool bars

The pull down menu has nine options that access most SPSS functions. These are:

File. With the file menu you can create a new file, open an existing one, or read in files created by other software (such as excel).

Edit. For modifying or copying text from output and syntax windows.

View. Allows you to select toolbars, change fonts, and view value labels in the data file.

Data. Makes global changes to data file, including split file, defining variables, and transposing variables.

Transform. Use this menu to make changes to specific variables or cases or to compute new variables from existing ones. You can also recode your data into the same variables or different ones. Changes made in this menu are temporary, unless specifically saved.

Statistics. From this menu you select the statistical procedure you want to perform. This includes frequencies, descriptive, t-tests, ANOVAs, and more.

Graphs. Creates bargraphs, histograms, scatterplots, etc.

Utilities. Use for miscellaneous functions, such as getting information about the data file.

Windows. Lets you switch between open SPSS windows.

Help. Allows you to receive help by topic, the SPSS web page, or use the SPSS Coach. The Coach suggests ways to analyse data, based on criteria you select.

The Data window is used to enter the data you will be analysing (More information on data entry for different tests are described with the tests). Columns represent variables (such as a subject's gender, score on a test, etc.) and rows represent subjects (such as the people completing your survey, or a television program you are coding). You can define your variable by double clicking on the column heading. This will open the dialogue box shown in figure 3:

>
figure 3: the define variable dialogues.

Enter the variable name at the prompt. The name cannot be longer with 8 characters and cannot contain spaces or symbols. This dialogue also lets you enter variable and value labels. To add these (and it is a good idea for many variables) click on the labels button. This opens the labels dialogue (shown beside the variable dialogue in figure 3). Here you can give the variable an alternate name (which can be longer than 8 characters), and give a label to each value in the variable. Type the value (number) in the value box, and the label in the value label box. Then press add. Other buttons on the define variable screen allow you to indicate the type of variable (text, numeric, etc) What to do with Missing Data and the Column Format. This allows you to right, left, or center justify your column.

When your data are entered, use the mouse to click on File on the pull down menu. Click on Save or Save as to save your data.
REMEMBER TO SAVE YOUR DATA FREQUENTLY!
If an error were to occur in the program (and it sometimes does) you lose all unsaved changes. This can be very upsetting, as you will have to start over again.

figure 4: the options dialogue.

At times it is necessary to modify your data set after it has been entered. You may want to create new variables, reverse the values in some of your variables, or analyse only a portion of your data.

figure 5: recoding data.

The Recode into different variables screen lists all of your variables on the left-hand side. Select the one you wish to recode, it will be copied into the middle column entitled Numeric variable --> Output variable. On the far right there is a box for you to create a new variable to recode values. Enter a name, then press Change, this will move that variable into the middle box. You have two choices for recoding variables, you can use the If or the Old and New Values. If allows you to reduce the number of values in a variable by numeric formula. The Old and New Values dialogue is shown. This allows you to change individual values in a variable. For instance, if you wanted to recode a 5-value variable, you would enter 5 in the old value box, and 1 in the new value box (because 5,4,3,2,1 would recode to 1,2,3,4,5). Then select add to add it to the menu. When completed select Continue and OK on the main dialogue.

figure 6: Split File.

Like the recode screen, the split file dialogue lists all of your variables on the left-hand side. Select Organize output by groups, then from the variable list, select the grouping variable(s). Click on OK. All analyses you do will be split by the grouping variable(s). This will remain in effect for all analyses until it is turned off. To turn split file off, open the same dialogue box and select Analyse all cases. Split file status is indicated at the bottom of the SPSS application window.

For most any study you conduct, you will need to gather descriptive information about your data. You do this using the Frequencies and the Descriptives dialogues. Both are located in the Statistics menu under Summarize. The Frequencies dialogue allows you to print out frequency counts, histograms, and descriptives such as means, medians, and standard deviations. The Descriptives dialogue allows you to print out the mean, standard deviation, minimum, maximum, and other descriptives. The Frequencies dialogue is shown in figure 7:

figure 7: Frequencies.

Select the variables for which you want frequencies. If you select nothing else, your output will contain freqeuncy tables for each variable. The Statistics button allows you to select other statistics for each variable, such as the mean and median. The Charts button allows you to select a graphical representation of each variable, such as a histogram. The Format button allows you to change the way the variable is displayed, including an option of sorting values alphabetically by label, in ascending numerical order, or in decending numerical order.

Charts and Graphs

SPSS will create charts and graphs to help you display your results. All graphs are available under the Chart menu. If you are using SPSS 8, your graphs will appear right in your output file. If you are using an older version of SPSS, your graph will be displayed in a chart carousel, and will have to be saved seperately. Figure 8 shows a bar chart made in SPSS.

figure 8: An SPSS Chart.

You can conduct both Pearson's or Spearman's correlations using the same dialogue. This is accessed through Statistics, Correlate, Bivariate. As seen in Figure 9, the default is Pearson's, but you can calculate Spearman's by clicking the checkbox beside its name. You can also choose if you want the significance level to be one-tailed (directional) or two-tailed (non-directional). If you want means and standard deviations, click the Options button, and click the Means and Standard Deviation check box and click Continue. Click OK to run the correlations.

figure 9: Correlation Dialogue.

Between-Subjects (independent) t-tests

figure 10: Independent T-test data.

To conduct the t-test go to the Statistics menu, select Compare Means and then Independent Samples T-Test. This opens the independent t-test dialogue, shown in figure 11:

figure 11: Independent T-test dialogue.

Like other statistics, all variables in the data file appear in the left-hand box. Find which variable(s) is the dependent one, and move it to the test variable box. In the figure, love kids [fem9] and tender [fem8] are the dependent variables. The grouping variable is the independent variable. It contains the groups you want to compare. In the figure it is gender. When the variable is first moved to the grouping variable box, it will appear as gender(?,?). You will have to click on the Define groups. . . button. This identifies the 2 groups you want to compare, in this example, male and female. Enter the corresponding value in the group 1 and group 2 box (1 and 2, respectively). When the values are entered, click on Continue. To run the t-test, click on OK.

Within-Subjects (paired) t-tests

figure 12: Paired T-test Data.

To conduct the t-test go to the Statistics menu, select Compare Means and then Paired Samples T-Test. This opens the paired t-test dialogue, shown in figure 13:

figure 13: Paired T-test Screen.

The variables from your data file will appear in the left box. To select the pairs of variables, the first variable highlighted will appear as variable 1 in the current selections box, and the second selected as variable 2. Click on the arrow to move the current selections to the paired variables box. In figure 13, fem2 and fem3 have been selected as pairs for a t-test. fem3 has also been selected for a second t-test. To complete this selection, you would select another variable. It would appear as variable 2. To run the t-test, click on OK.

One Way ANOVA

One step up from the t-test is the one-way ANOVA. Use this if you have one independent variable with more than two levels. To access the one-way screen, go to Statistics, Compare Means and then One-Way ANOVA. A sample screen is shown in figure 14:

figure 14: One-Way ANOVA.

Select from the variable list one or more dependent variables, and move them to the Dependent Variable List. Select your independent variable and move it to the Factor box. If you want means and standard deviations for each cell, click on the Options button. If you want to conduct a post hoc test (to find where differences lie) click on the Post Hoc button. When you are satisfied with what you have selected, click on OK to run the ANOVA.

General Factorial ANOVA

If you have two independent variables (both between subjects) you use a General Factorial ANOVA to analyse the data. Go to Statistics, General Linear Model, GLM - General Factorial. This will open the dialogue shown in figure 15:

figure 15: General Factorial ANOVA.

Select your dependent variable from the variable list and move it to the Dependent Variable box. Then select your independent variables and move them to the Fixed Factors box. You do not have to specify the range, SPSS will just use all values in each variable for the range. Use the Options button if you want Means and Standard Deviations for each cell, and the Post Hoc button if you want to conduct post hoc tests. Click OK to run the ANOVA.

Repeated Measures ANOVA

If there are any within subject factors in the research design, you must use a repeated measures ANOVA to analyse the data. Go to Statistics, General Linear Model, GLM - Repeated Measures. This will open a repeated measures define factors box. This identifies factors that your variables comprise. for the data in figure 10, the dependent measures are time and mistakes. In the define factors box, shown in figure 16, fill in the within-subject factor name with a logical factor name, in the figure it is named test. Also, you must fill in the numbers of levels. there are two levels in the figure, one for the variable time and another for the variable mistakes.

figure 16: Factor Definition.

Click Add to move the factor to the middle box. Then click Define. This opens the repeated measures ANOVA dialogue box, shown in figure 17:

figure 17: repeated measures ANOVA.

This is where you identify the variables comprising your factors. As seen above, the factor "test" was said to have two levels, thereby the _?_(1) and _?_(2). Highlight the appropriate variables and move them to the right box to identify the 2 levels. If you have between subject variables (such as gender in figure 17), move them to the Between Subjects Factors box, and then define the range by clicking the Define Range. . button, the same as you would for a t-test analysis.

More on Chi-Squares will be added soon

For the regression command choose Statistics, Regression, and then Linear to open the linear regression dialogue box. Your variables will appear in the left hand box, as shown in figure 18. Click on the appropriate [predictor and criterion] variables and transfer them tothe appropriate boxes [dependent and independent] by clicking on the arrows. The default method is "enter".

For Multiple Regression the same box is used. The default is "enter" where all the available independent variables are entered into the equation directly. Other choices are : stepwise: independent variables are added to (or removed from) the equation one at a time, the order being determined by statistical considerations. Forward stepwise, regressors are added to the equation one at a time. Backward elimination method, regressors are subtracted one at a time.

figure 18: Linear Regression Dialogue.

Perfection is hard, so if there are mistakes in this document, please let me know. Also if you have comments on other things you would like to see about SPSS, please send an email to The TA.
Last modified on February 15, 1999.