Before you dig into statistical analysis, it’s a good idea to first get a graphical view of your data. This simple step can alert you to any serious problems that would undermine all your attempts to find statistical significance.

In the menu of SPSS, select Graphs and then Chart Builder. You should get the following alert:

As mentioned on the last page, you need to be sure to set the "Measure" correctly for each variable. In particular, be sure to identify your dependent variable as "Scale" if it is on an interval or ratio scale. You have already done that, so go ahead and click "OK".

Checking Distributions

Many statistical analyses assume that the dependent variable is approximately normally distributed, which means that most of the scores are clustered around the mean and there are approximately equal numbers of scores above and below the mean. A good plot to use to investigate that assumption is the Box Plot. To get started, select "Boxplot" from the "Choose from" list in the lower left side of the dialog, then pick the first example on the left ("Simple Boxplot") and drag it up into the Preview window:

Now you need to identify what will go on the x-axis (horizontal, along the bottom of the graph) and what will go on the y-axis (vertical). Drag "Condition" from the "Variables" area in the upper left of the dialog down to the "X-axis?" area. Drag "Tip_Percentage" onto the "Y-axis?" area. You should get this:

Note that the example that SPSS gives you in the Chart Preview is a complete fiction. It shows three categories along the x-axis, but your data only has two. I’m not sure why they can’t draw an example of the real plot you will get, but for whatever reason, all you see in the Preview is an idea of the final plot. Click "OK" to make the plot. You should get the following in the Output window:

Interpreting a Boxplot

A boxplot is designed to graphically show you several pieces of information about the data at the same time:

  1. The shaded box in the middle shows you the boundaries of the middle 50% of your data. It is called the "Interquartile Range". The bottom 25%, or "quartile", is below the box, and the upper quartile is above the box.
  2. The horizontal line inside the shaded box is the median. Exactly half of the data is above it and half is below it.
  3. The whiskers extending above and below the box go out to EITHER the minimum and maximum values OR to 1.5 times the length of the box (1.5 times the Interquartile Range, or IQR), whichever is closer to the median.
  4. Sometimes (not in the present data), there will be data points outside the whiskers. These are worth looking at closely because they could be "Outliers", data points far from the center that may cause misleading results if you rely on the mean. SPSS has different markers depending on how extreme the scores are. If they are marked by circles, they are within 1.5 IQRs of the median. If they are marked by asterisks, they are between 1.5 and 3 IQRs.

This plot suggests that chocolate is having an effect: the tip percentage appears to be slightly higher in the Chocolate condition than in the No Chocolate condition. It also suggests that there is more variability in the Chocolate condition (the IQR and whiskers are wider).

Getting rid of the gray background

SPSS believes in using gray backgrounds to ease eyestrain when reading plots on a screen. Unfortunately, this can make it harder to read a plot when printing it on paper. To get rid of the gray background, double-click on the boxplot in the Output Window. A Chart Editor dialog opens. Now, double-click on the gray background. Another dialog opens called "Properties." Just above the multi-color palette are three color rectangles: black, white, and white with a red line through it (transparency). Click on the white rectangle, which should make the "Preview" box white:

That should give you a nice white background on the plot:

On the next page, we will proceed to testing whether the difference between the chocolate and no-chocolate groups is statistically significant.