Create a two-way table with SPSS

In 1999 the news media ran a number of stories about a recently released research study which suggested that using a nightlight during a child's first years may lead to nearsightedness. Researchers from University of Pennsylvania and Children's Hospital of Philadelphia collaborated on a study of 479 children as to the lighting conditions under which they slept between birth and 2 years of age (lamp, night light, or no light) and whether or not they developed nearsightedness. The researchers were interested in the relationship between a young child's nighttime exposure to light and his/her later nearsightedness.

The dataset contains 479 observations and two variables:

• Light: lamp, night light, no light
• Nearsightedness: no, yes

Dataset

• nightlight.xls
• an SPSS version of the dataset is available on your class website: nightlight.sav

Open the dataset in the SPSS data editor.

The following instructions are based on the student version of PASW (SPSS) version 18.

Create a two-way table of nightlight vs. nearsightedness

Before beginning the analysis, you will need to determine which variable is explanatory and which is response.

Explanatory vs. Response

• The nightlight variable is the explanatory variable.
• The nearsightedness variable is the response variable.

Use the Crosstabs (short for crosstabulation) routine in SPSS to create the two-way table.

• Select Analyze > Descriptive Statistics > Crosstabs....

The Crosstabs dialog box opens.

• Move the explanatory variable into the Column(s): field.
• Move the response variable into the Row(s): field.

To include conditional percents in the table:

• Select the Cells... button.

The Crosstabs: Cell Display dialog box opens.

• Check the box for Columns. (This will create percents conditioned on the column variable, which we set as the explanatory variable.)
• Click Continue.

The cells dialog closes.

• Click OK.

The Crosstabs dialog box closes and SPSS activates the Output window to display the two-way table.

Create a clustered bar graph

To further describe the relationship, consider creating a clustered bar graph showing the conditional percents.

• In Chart Builder, under the options for Bar, drag the Clustered Bar image to the preview window.
• Drag the response variable to the X-Axis? field.
• Drag the explanatory variable to the box in the upper right labeled Cluster on X: set color.

In the Element Properties dialog box:

• Click the down arrow under Statistic: to change the y-axis statistic from Count to Percentage(?).
• Click the Set Parameters button.

The Element Properties: Set Parameters dialog box displays. To specify a conditional percentage, in the Denominator for Computing Percentage field:

• Select Total for Each Legend Variable Category (same fill color).
• Click Continue.

In the Element Properties dialog:

• Click Apply.

In the Chart Builder dialog:

• Click OK.

The graph displays in the Output window. Use the Chart Editor to add a title, adjust the y-axis label to indicate percentage of lamp, night light, or no light, and add the percentage label to each bar (Elements > Show Data Labels).

Interpret the two-way table

Compare the distribution of percentages for nearsightedness among each of the sleeping conditions: a lamp, a nightlight, or no light. Use the conditional percentages to draw a conclusion as to the relationship between using a light during early childhood and later nearsightedness. Based on the results of this study, what might you suggest to parents?