Hi.
I need to calculate a total score with 8 cognitive test but TMT raw score have opposite direction compared to the other tests. I want all dimensions have the same direction in T score, in other words, lower values represent lower cognitive perfomance. How to change the direction in SPSS?
Note: TMT raw score is time used for to complete the test
Thanks!
Seconding Jorge: TMT (Trail Making) times are often very skewed, so using the inverse transformation (1/x) often normalizes the distribution while at the same time making high values better. It's even interpretable as speed (1/s) rather than time.
Combining the scores is a different issue. Rather than, say, summing normalized scores (z-scores) you really should do the recommended factor analysis. At the same time you can check whether the tests really measure the same thing to various degrees, or whether there's separable dimensions.
If you are only worried about reversing the time score, Luke's suggestion is OK. but I'm not sure if it is the only transformation you should apply in order to obtain your total. Maybe you need to use some factor analysis or other statistical method to properly define the total and not just add up the cognitive tests scores and the time (reversed).
Seconding Jorge: TMT (Trail Making) times are often very skewed, so using the inverse transformation (1/x) often normalizes the distribution while at the same time making high values better. It's even interpretable as speed (1/s) rather than time.
Combining the scores is a different issue. Rather than, say, summing normalized scores (z-scores) you really should do the recommended factor analysis. At the same time you can check whether the tests really measure the same thing to various degrees, or whether there's separable dimensions.
if you want to get all your T scores in the same direction (bad peformance - better performance) you should use 'compute variable" in SPSS and go with:
TMT_Tscore(your new variable) = ((mean - TMT raw score)/SD) * 10 + 50
I usually transform my raw score to T score with that formula instead of the regular form in test where a greater score means worst performance (test where the raw score = time spent; p.e. TMT, Grooved Pegboard Test, time spent in ROCF, ...)
i dont know if this is correct but if you want to reverse just raw scores maybe you could try substract the max score of your sample to every single raw score (TMT_rawtrans= MAX - TMTraw)
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Statistics
- Statistical Software
- SPSS