z-score help?

#1
Hi, I asked 100 people to make a single rating one a 40-point scale. Half of them got a scale ranging from 10-50. The other half got a scale from 60-100. These are organized as two columns of (incomplete) data, column A and B. I'd like to compare their means using an independent-samples t-test, but I need to put them on the same scale first. I thought I'd do this by computing z-scores. So in SPSS, I first generated z-scores for each value by going to Descriptives / Save standardized values. Then I combined those new z-scores into a single column and then ran the t-test on that. But the test came out saying t = .0000, p = 1.0000. So I'm assuming I did something wrong, but what could it be? The order of operations? The standard errors look unremarkable (0.12). Thanks!
 

katxt

Well-Known Member
#2
The standardization makes the mean = 0 for both groups so there won't be a difference in the standardized scores.
Perhaps you could try making them both 10-50 by taking 50 off each score in the 60-100 group.
 
#3
Ah that makes perfect sense, thank you!

I still welcome other comments if there are other conventional approaches, but this should work for me in the meantime :)

The standardization makes the mean = 0 for both groups so there won't be a difference in the standardized scores.
Perhaps you could try making them both 10-50 by taking 50 off each score in the 60-100 group.