Hi,
I am looking to do a comparison between results of my primary data study on nutrition outcomes with data from a standard health survey (NFHS). My study was done for only malnourished children so I am drawing out SAM cases from the raw NFHS data set for comparison.. I want to see whether the differences are significant on different indicators (around 10 indicators common in both data sets)..
Please suggest appropriate methods to analyse whether differences on certain parameters between these 2 studies are significant..(suggest references)?
Marian,
You can go for Z-test to compare mean of two group of variables in case of large no. of data set. If the data set is less than 30 then you can do independent sample t-test for the same. You can use SPSS (any version) to perform these tests. However, if you are not aware of SPSS, then try it in MS Excel of Windows. For that you just go to data tab, then data analysis section and the desired test you want to perform.
Thanks
Convert to generic effect size and conduct Fisher's r-z transformation with chi-square df=1 (Dif between 2 samples) is significant.
Hope this helps,
Matthew
Logistic regression may work for your case, by which you can adjust for co-variables confounding your observed associations.
If the data was numerik and normally distribution you can use the independent t test (sample < 30) or z test if the sample is large.
If the data was not normally distribution you can use the mann whitney test. If the data was ordinal it can be tested by chi square test.
- Badges
- Science topic
- Similar topics
- Medicine
- Epi
- Age Groups
- Children