Selecting a Post Hoc Test after ANOVA in SPSS

This video demonstrates selecting the appropriate post hoc test after ANOVA in SPSS. Included are discussions of post hoc tests such as LSD, Tukey, Bonferroni, REGWQ, Hochberg’s GT2, and Gabriel.

[MarkVanderley]

This is helpful in understanding how to use a post hoc and what things to consider when choosing one

[adamwilson]

It is interesting to me how many different post hoc tests there are. It makes me wonder if they are all necessary or if it is just competition among statisticians.

[grishp]

Hello Dr. Todd Grande, do you have a separate video for analyzing the 'Two-way ANOVA data when the levene's test shows significant value'? I watched your video when the levene's test shows significant value for one-way ANOVA (you had performed the Games-Howell's test) in that case. However, I'm really really interested in knowing what happens when it's two-way ANOVA. Thanks for your videos.

[vanessagraham]

Yeah, some assistance with defining "large" and "small" groups and "large" and "small" differences between groups would be helpful.

[ivybyun]

The selections of options was quite speedy, so I will need to rewatch this, but I appreciated the explanation of the different nuances amongst post hoc tests.

[corieschoeneberg]

Can you provide more specifics regarding what is considered "large" and "small" for group sizes?

[thomasstarr]

Makes a lot of sense and seems to hold together nicely in terms of processing the data results. When you were generating the plots are there guidelines to follow? How did you know to use multiple variables multiple times?

[elizabethadair]

I used the Scheffe's post hoc test and the Bonferroni. The Scheffe's test had a more conservative result than the Bonferroni for the sample data I ran.

[chaiyongj]

Thanks for good explanation. In case of 2 way ANOVA analysis with 2 x 2 design (genders- male, female*being athlete-yes, no), I tried to analyse the post hoc test but the result was ignored because there are fewer than 3 levels for each independent variable. Can I tried the post hoc test via the 1-way ANOVA analysis? It analyzed the multiple comparisons across 4 groups. Is that wrong? What should I do? Or I go to use independent t-test for each comparison?

[Mina]

Sir, can we use Tukey if the sample size of group 1 is 40, group 2 is 10 and group 3 is 4... which one will be appropriate in this situation?

[taylorjohnson]

Why would a post hoc analysis not be able to run when an independent variable only has 2 levels?

[lishab]

What happens when the Levine's test is more than 0.05?

[haiuuyen]

Thank you, Bro! This video really helpful for my project!!!

[athena]

could you explain the reasons why you choose those stuff?

[MiguelSMunoz-vmgr]

Thanks man ! this video really helped me!