An article I came across, which was an analysis of two prospective study cohorts (one male cohort and one female cohort) provided a pooled results of the combined cohorts for the particular exposure-effect relationship. Can someone tell me under what assumptions a pooled analysis is undertaken and what are the advantages?
As far as I understand, it means combining results and then analyse them. You can say Meta analysis. Generally meta-analysis is used to increase the statistical power of the analysis.
In this case, estimates are obtained from each gender and then combined. This is one way of removing the gender effect when estimating a single estimate of exposure effect. When the exposure effects from the genders are combined, it is assumed that the exposure effects are the same in the two genders. Therefore, the final exposure effect is a within gender estimate. From a modelling perspective, outcome is the dependent variable, exposure is the factor of interest, and gender is the covariate to be adjusted for or to be controlled of.
The oft fogotten danger is that you run the risk of Simpson's paradox, so pooled data may mislead you.
Pooling of results is a Meta-analysis method used to combine the results of different studies in order to get qualitative analysis. Usually used when the size of study is too small to evaluate the effect or relationship. So, pooling results will increase the power of statistical analyses.
You can use a multiple imputation to complete data set. Pooled output can be used if there are missing values or you want more accurate results. According to SPSS pooled results are generally more accurate than those provided by single imputation methods
This article may be of interest too: http://ehp.sagepub.com/content/24/2/218.abstract
What are the conditions of a Pooled analysis? how can we use it on SPSS?
- Badges
- Science topic
- Similar topics
- Medicine
- Physiology
- Biosignals
- Biomedical Signal Processing
- Statistical Analysis