While testing data for eg. through repeated t-test or ANOVA, is it necessary that both groups should be measured on same scale. if one group is measured on 11 point scale and other on 7 point scale.Can we apply test directly or Is there any need to bring them on same scale.
Mh, if one variable is measured in percent of agreement (0-100%) and the other one is measured as probability (0-1), would you say that the variables should be converted, although both measure the same content?
Apparently 'yes'. Same for your case, imagine all participants in both groups would totally agree (very artificial and would not work because there is not within group variability, but thats not the point) but the first group has maximum of 7 the other 11, can you compare this? It would seem as if there is a huge difference of 4, but this is only due to the scaling.
Should have been necessary, because the average between the different data measurement scale is different, so if you want to compare it, it should normalize the data using a data normalization techniques (xi - minimum xi)/range data or standardize using z scores (xi - xbar)/stdev
Jagdish, What are you looking for by comparing the averages? Look at these distributions. All they have the same mean. Averages comparison is not a good idea to analize the differences in the two data sets.
- Badges
- Science method
- Similar topics
- Mathematics
- Statistics
- Statistical Modeling
- Analyses of Variance
- ANOVA