# Thread: When/ when not to use Z scores to compare data

1. ## When/ when not to use Z scores to compare data

In comparing 2 groups of numbers, could you give me any examples of when would it be (1) beneficial (2) when it would not be beneficial to first convert each group of numbers to z scores?

In comparing 2 groups of numbers, I have been asked to consider the following:

1) Center
3) Shape
4) Outliers

in determining whether it would be:

BENEFICIAL to first convert each group of numbers to z scores
NOT BENEFICIAL to first convert each group of numbers to z scores
NOT HAVE AN EFFECT ON THE COMPARISON to first convert each group of numbers to z scores

Any thoughts on this? I have looked at textbooks/ searched the internet but am having difficulties getting explanations of when/ when not to convert to z scores for purposes of comparison in relation to each of the above situations (center, spread, shape, outliers). Your input on any/ all of the above would be much appreciated. Thanks in advance.

NB. This question does not involve any specific data.

3. ## Re: When/ when not to use Z scores to compare data

Well you lose all information about the center and spread when you convert to z scores so it would be harmful to convert to zscores to assess those quantities. You would retain the same shape after converting but it wouldn't really 'help' you in any way. Converting to z-scores could potentially help identify outliers though.

4. ## Re: When/ when not to use Z scores to compare data

Originally Posted by Dason
Well you lose all information about the center and spread when you convert to z scores so it would be harmful to convert to zscores to assess those quantities. You would retain the same shape after converting but it wouldn't really 'help' you in any way. Converting to z-scores could potentially help identify outliers though.
Thanks for this information; this helps a lot.

5. ## Re: When/ when not to use Z scores to compare data

Originally Posted by Dason
Well you lose all information about the center and spread when you convert to z scores so it would be harmful to convert to zscores to assess those quantities. You would retain the same shape after converting but it wouldn't really 'help' you in any way. Converting to z-scores could potentially help identify outliers though.
Hi Dason, I was wondering if you/anyone else could help me on this follow up question. As you stated, converting to z-scores could potentially help identify outliers. At the same time, I also was under the impression that because outliers affect the mean and standard deviation, they in turn can affect the z-scores. Does this mean that the presence of outliers will cause problems with the comparison of datasets after converting to z scores? Basically I am confused as to whether outliers would facilitate the comparison of 2 datasets after converting to z scores OR hamper comparison.

 Tweet

#### Posting Permissions

• You may not post new threads
• You may not post replies
• You may not post attachments
• You may not edit your posts