SPSS Statistics

 View Only
Expand all | Collapse all

#SPSS proportional meta analysis

  • 1.  #SPSS proportional meta analysis

    Posted Thu May 04, 2023 09:40 AM


    Does SPSS have the capability to undertake a proportional metaanalysis and if so how?
    I have several studies that record percentage of participants achieving an outcome after a medical treatment, but with no control group. I can't see how to carry out this type of meta analysis on SPSS. 

    Many thanks, 


    Charlotte Gollins

  • 2.  RE: #SPSS proportional meta analysis

    Posted Thu May 04, 2023 11:18 AM

    Hi. I talked to one of the statisticians, and got this reply:

    "The meta-analysis procedures were not designed to include the feature of proportional meta-analysis in an automatic way.

    "There is a way that the customer may want to have a try to see if it meets their needs. For dichotomous data, first convert them into percentages. Then, treat these percentages as pre-calculated effect sizes, and use Meta Analysis -> Continuous Outcomes -> Pre-calculated Effect Size to analyze the data.

    "I believe it should return meaningful pooled results. However, they should be cautious about interpreting the standard errors and confidence intervals. They may not be what he/she wants. After all, the percentages are treated as a continuous effect-size variable."

    Rick Marcantonio
    Quality Assurance