I noticed that the median and minimum, maximum obtained from two different commands in SPSS 25, explore and frequency commands, are different from each other. Is it because "explore" deletes outliers or is there a problem?
P.S: my data is not normal and contains many outliers.
That's very concerning. Can you see if this happens with any other dataset? The median, minimum and maximum should be the same no matter what.
Dear Soroush,
I think the name of the second command for generating summary statistics is Descriptives, not Frequency. I don't know whether this is anything to do with your anomaly.
Soroush, can you provide a small data set that can be used to duplicate the problem?
Meanwhile, have you read carefully the documentation for the two procedures you are using?
- https://www.ibm.com/support/knowledgecenter/en/SSLVMB_25.0.0/statistics_reference_project_ddita/statistics_reference_project_ddita-gentopic2.html
Given that you've said the min and max values differ from the two procedures, my first guess would be that you are handing each procedure a list of 2 or more variables, that there is some missing data, and that one of the procedures uses listwise deletion by default and the other does not.
If that turns out to be the problem, and you get min and max values that match, it would not totally surprise me if the medians still differed by a small amount. IIRC, EXAMINE and FREQUENCIES may use different algorithms for the median. (Again, consult the documentation.)
@ Peter: The DESCRIPTIVES command does not report the median. Note too that one can suppress the frequency table from FREQUENCIES when using it to compute a median. E.g.,
- https://developer.ibm.com/answers/questions/219099/getting-a-median-in-descriptive-statistics-in-spss/
HTH.
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Statistics
- Statistical Software
- SPSS