I was wondering if anyone knew how to perform the Haldane correction for SPSS (this is the definition of the correction - https://www.oxfordreference.com/view/10.1093/acref/9780199976720.001.0001/acref-9780199976720-e-1977).
I have some zero cell values for several adjusted risk ratio tests I am performing, but I don't know how to implement this correction on SPSS.
I don't think I can manually add 0.5 to all cells because of the adjustment for age and sex within the calculation.
Any help would be much appreciated.
Best wishes
Sam
Hello Samuel,
You can enter data by cell/subgroup, and use any number (e.g., observed frequency + 0.5) as a weight in spss. (Data/Weight cases/Weight cases by...) The only problem is that not all subprograms in spss allow for non-integer cell counts (for example, crosstabs w/ chi-square test of independence rounds 0.5 fractions up to the next higher integer). That would "over-correct" in some subprograms.
You didn't indicate what specific analysis you were running, or how many control/covariates were in your model, so it's hard to advise whether this would all be simpler if done using a spreadsheet, for example.
Good luck with your work.
- Badges
- Science topic
- Similar topics
- Correction