HI. I have been looking for answer for the question above. I'm trying to reproduce a study odds ratio. I've attached the sample table i would like to create. What I'm not able to find out is how do i calculate odds ratio in SPSS with with one of categories in the variable set at 1.0? If you see the table ie the Charlson Morbidity score of 0 has odds ratio of 1.0. Same with Resection, Duke's A and Elective. How do i set that in SPSS? Help appreciated.
The rows that show OR = 1.00 and have no confidence interval reported are the reference categories for the ORs in the other categories. E.g., for Charlson Morbidity, the odds of the outcome occurring are 1.48 times greater for folks with a score of 1 relative to those with a score of 0; and the odds of the outcome occurring are 1.69 times greater for those with a score of 2 relative to those with a score of 0.
You have not provided enough info to allow one to reproduce the odds ratios in SPSS (or any other package). To get crude odds ratios (i.e., not adjusted for other variables), you would have to provide the contingency tables for explanatory variable x outcome. With that info, you could use WEIGHT CASES to weight the variables by the cell counts, and then use the LOGISTIC REGRESSION command (or GENLIN with binomial error and logit link, etc.)
HTH.
Hi Bruce Weaver . Thank you for you answer. How do I put the reference category in spss to calculate OR of the variable? And what do you mean I don’t provide enough Info? sorry for more questions but this is the thing stopping me from producing my report.
See if this example helps.
* --- Start of syntax ---.
* Example taken from this website:
* https://libguides.library.kent.edu/SPSS/WeightCases.
NEW FILE.
DATASET CLOSE ALL.
* Read in the 6 cell counts.
DATA LIST LIST / Rank PickedMajor n (3F5.0).
BEGIN DATA
1 0 212
1 1 114
2 0 171
2 1 168
3 0 92
3 1 198
END DATA.
VALUE LABELS
Rank 1 "Freshman" 2 "Sophomore" 3 "Junior" /
PickedMajor 1 "Yes" 0 "No".
* Weight the cases by N.
WEIGHT BY N.
CROSSTABS Rank BY PickedMajor.
* Estimate a logistic regression model.
* Use the first category of Rank as the reference category for ORs.
LOGISTIC REGRESSION VARIABLES PickedMajor
/METHOD=ENTER Rank
/CONTRAST (Rank)=Indicator(1)
/PRINT=CI(95)
/CRITERIA=PIN(0.05) POUT(0.10) ITERATE(20) CUT(0.5).
* Repeat, but with the last category as the referent.
LOGISTIC REGRESSION VARIABLES PickedMajor
/METHOD=ENTER Rank
/CONTRAST (Rank)=Indicator
/PRINT=CI(95)
/CRITERIA=PIN(0.05) POUT(0.10) ITERATE(20) CUT(0.5).
* --- End of syntax ---.
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Statistics
- Statistical Software
- SPSS