Using SPSS in Independent Sample t test can I create more than two group? I am attaching the snapshot, I cannot find third group option. Is there any other test?
Safa Aouinti Thanks. Yes for ANOVA, as per my understanding I need to have independent & dependent variables.
I just wanted to set one group as benchmark & compare with three others. & Independent sample t-test to me is very easy to explain for corporate executives (As they are least interested to hear statistical things).
Thank you for your response. I divided data into two groups and presented it with sample t-test results.
If you want to compare several groups to one --- and assumptions of the test are met --- you could use Dunnett's test. You can present the results very simply with a plot of the means, and stars or letters indicating which treatments differ from the benchmark. Something like this: https://rcompanion.org/rcompanion/images/d_05_05.jpg . You can't get simpler than that in presenting results.
In general, you shouldn't use a series of independent t-tests, because they ignore the overall variability of the data set as a whole.
I second Salvatore S. Mangiafico's recommendation of Dunnett's test, but will add that the reference (or control) group must be specified in advance. In other words, you can't wait until you've seen the data to identify the reference group. Bear in mind too that Dunnett's test can be carried out even if the omnibus F-test (in the ANOVA table) is not statistically significant. Among the most common multiple comparison methods, only Fisher's LSD requires a significant omnibus F-test. HTH.
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Statistics
- Statistical Software
- SPSS