This page will give you an introduction to analyzing data, getting descriptive statistics, and creating histograms to check assumptions of normality.
All analyses in SPSS are located under the "Analyze" menu. Each listing under the "Analyze" menu has several sublistings and it is easy to feel intimidated by the long list. We will walk through each common analyses one by one.
After selecting an analysis, SPSS will provide you with a window in which to select the variables and specifications for your analysis. Once you have set these parameters, SPSS will run the analysis and open a new window (Figure 2.1) called the "SPSS Viewer" (as opposed to the "SPSS Data Editor" in which we have been working in so far) where you may view the output of your analysis. All analyses will be placed in the same SPSS Viewer window unless the viewer is closed and all previous analyses saved separately. We will discuss using the viewer in more detail here, but for now we will discuss how to properly conduct several common analyses.
Figure 2.1 The SPSS Viewer
No matter what type of analysis you intend to run, you will almost always want to include the Means, Standard Deviations, and other basic information about your data set in your report.
1) Click on "Descriptive Statistics" in the "Analyze" menu.
2) Choose the option "Descriptives..." from this menu.
3) From the "Descriptives" window (Figure 2.2), choose which variables for which you would like descriptive statistics.
Figure 2.2 Descriptives Window
4) By pressing "Options" on this screen (Figure 2.3), you can change which descriptive statistics will appear in the SPSS viewer, shown in Figure 2.4.
Figure 2.3 Options for Descriptive Statistics
Figure 2.4 Basic Output for Descriptives
You can also use the "Descriptive Statistics" option, under the "Analyze" menu, to access frequency information for each variable. Simply follow the steps above, but select "Frequencies..." rather than "Descriptives..." From the "Frequencies" Window (Figure 2.5), you can choose different graphic representations for your data (such as histograms) by pressing "Charts."
Figure 2.5 Frequencies Window
Figure 2.6 Basic Output for Frequencies
There are many other ways to explore your data in SPSS. The "Explore" option will give you many descriptive statistics, stem-and-leaf plots, boxplots and histograms all in one go.
1) Click "Descriptive Statistics" under the "Analyze" menu.
2) From this menu, select "Explore..."
3) From the "Explore" window, select a dependent variable. If you would like to split analyses by another variable, put that variable in "Factor." Here, by putting "gender" as a factor, it will give separate descriptive statistics and charts for males and females in their answers to "satisfaction with major."
Figure 2.7 Explore window
4) Click on "Plots."
5) As seen in Figure 2.8, select "Stem-and-leaf" and "Histogram" under "Descriptive" and press "Continue."
Figure 2.8 Plots menu: stem-and-leaf and histogram selection
6) From the "Explore" window, press "OK" to view output of your selected variable.
Would you like your histogram to display a normal curve as well?
1) Select your histogram and double-click on it.
2) A new window will appear, with different menus at the top of the screen, as seen in Figure 2.9. This is the chart editor. The histogram on your regular output screen will stay shaded until you close the chart editor.
Figure 2.9 Chart editor window
3) Go to the "Chart" menu, then select "Options..."
4) Click on "Normal Curve" as shown in Figure 2.10. You should now have a normal curve superimposed over your histogram.
Figure 2.10 Histogram options window
5) To go back to your output screen, close the chart editor.
To alter the appearance of the chart, see the section on making graphs.
Making histograms is one of the best ways to check your data for normality. Histograms will only display values that are present in your dataset; for example, if you have a 1-9 scale and nobody chooses 9, then 9 will not appear in your histogram. Many analyses require some degree of normality; you can look at your data and decide if it is approximately normal. Choosing the "Explore" option will generate a histogram for you if you ask it to, but you can also make histograms separately.
1. From the Graphs menu, select "Histogram."
2. Put your variable in the "variable" box.
3. To see the normal curve superimposed on your histogram, click "Display normal curve" as shown in Figure 2.11.
Figure 2.11 Creating a basic histogram
4. Click "OK."You will see the histogram with a legend displaying standard deviation, mean, and N. To alter the appearance of the histogram, see the section on making graphs.
Sometimes, such as in a factorial ANOVA, you may want to make histograms of your data split by two or more factors. Here, we will make histograms of satisfaction with major split by school (Reed, PSU, U of O) and by gender (male, female) using an interactive graph.
1) Under the Graphs menu, select "Interactive."
2) Select "Histogram..."
3) You will see a screen like that shown in Figure 2.12. You will see a single small box with an arrow through it below and to the right of a box that says Count[$count]. Drag your dependent variable (here, satisfaction with major) into this box.
4) Drag your independent variables (here, gender and school) into the large box that says Panel Variables.
Figure 2.12 Creating a histogram split by two factors
5) To put normal curves on each of these, click the tab on the top of the window that says "Histogram." Click on the box that says "Normal Curve."
6) Click OK to create your histograms.