Z-scores with non-normal data


I am reviewing someone's study in which they used z-scores to standardize and compare data. I see no mention of their test for normality and when I looked at their raw data, I found much of it to be heavily skewed and not-normally distributed.

I need to determine whether this means the whole thing is junk, or if it is still a useful analysis, albeit not as valid as it would be if they had transformed the data (i.e., can/should I call it junk or is that an overstatement). Unfortunately, I do not have time to transform and re-run myself.



Less is more. Stay pure. Stay poor.
How did they "compare" it?

You are right in some of your assumptions. If you transformed skewed data it will still be skewed even though it is standardized. You can transform it, which could take relatively no time at all and run a normality test.

The big issue would be whether or not after they transformed it they used the right comparisions and met its assumptions.