One of my colleagues has asked me for help with a SPSS problem that left me confused. When perfoming an ANOVA using the Scheffé's method to adjust the significance levels, we get different significance values with different SPSS versions using the same set of data. For example, the output of SPSS 10.1 shows a significance value of 0.057 while the output of SPSS 11.0 shows a significance value of 0.029. All other values (mean differences, standard errors, confidence intervals) are exactly identical. Tests with newer SPSS versions up to v16 revealed the same problem, whereas only the significance value varies. I rifled through various SPSS forums serarching for an explanation but found nothing.
Any ideas on why we get different significance values with different versions of SPSS?
Addendum: I have uploaded an example for the differences on my ScienceFeed account:
http://www.sciencefeed.com/media/image/message/22a3e0f3298bbe455d2e72cf30bbdddb
The later versions should be the more appropriate value. This is because the confidence interval does not include 0, which would be the case for 0.029, but not for 0.057.
G
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics