StATS: Inputting a two-by-two table into SPSS (created 1999-09-18)
Dear Professor Mean, I have the following data in a two by two table:
When I try to enter this data into SPSS, I can't get it to compute risk ratios and confidence intervals. What am I doing wrong? -- Jinxed Jason
You have values ranging from F- to D+? I hope this isn't data on the grades you received in college.
Actually these data are from a paper: Sands et al (1999). F+ represents presence of a risk factor (in this case, previous miscarriage) and F- represents absence of that risk factor. D+ represents presence of a defect (ventricular septal defect or VSD) and D- represents absence of that defect.
1.3 (0.7,2.3) Female VSD
2.1 (1.3,3.2) Low parity VSD
1.1 (0.7,1.8) Smoking VSD
0.8 (0.5,1.3) Alcohol VSD
Notice that we have to do a bit of arithmetic to get all the values. If 34 out of 173 VSD cases had a previous miscarriage, then 139=173-34 did not. If 23 out of 142 controls had previous miscarriage as a risk factor, then 119 did not.
For data like this, you have to re-arrange things and then apply weights. The following discussion talks about SPSS, but the general method works for most other statistical software.
To re-arrange the data, you need to specify three variables: F, D, and COUNT. F takes the value of 1 for F+ and 0 for F-. D takes the value of 1 for D+ and 0 for D-. The 0-1 coding has some nice mathematical properties, but you could use 1 and 2 instead. For each combination of F and D we will record the sample size in COUNT.
Here's what your re-arranged data would look like
Enter the data, and tell SPSS that W represents a weighting variable, and you're ready to rock and roll. You do this by selecting Data | Weight Cases from the SPSS menu.
Then select Analyze | Descriptive Statistics | Crosstabs from the SPSS menu to create a two by two table.
Be sure to click on the Statistics button and select the Risk option box to ask SPSS to compute the risk ratios.
I also usually find it useful to display the row percentages. To do this, click on the Cells button.
In the Crosstabs: Cell Display dialog box, select the Row Percentages option box.
Here's what the first part of the output looks like.
Notice that the rows and columns are reversed in this table. There are several ways to change how the table is displayed, but it is showing essentially the same information in any order.
Here is what the second part of the output looks like.
By the way, if you tried to use the crosstabs procedure without weighting, you would get exactly one observation in each cell. Pretty boring, eh?
Jinxed Jason can't figure out how to enter data from a two by two table into SPSS. Professor Mean explains that you need three variables to represent a two by two table. The first variable indicates the specific column of your table and the second variable indicates the specific row (or vice versa). The third variable indicates the count or frequency for each intersection of row and column. You do not include the row or column totals in your data entry. You can then select Analyze | Descriptive Statistics | Crosstabs from the SPSS menu to analyze the data from your two by two table. You get additional analyses by selecting the Risk and/or Chi-square option boxes.
This page was written by Steve Simon while working at Children's Mercy Hospital. Although I do not hold the copyright for this material, I am reproducing it here as a service, as it is no longer available on the Children's Mercy Hospital website. Need more information? I have a page with general help resources. You can also browse for pages similar to this one at Category: Data management or Category: SPSS software.