What is sphericity? Explained Simply

[@jellyfisheyes]

Truly, angels live among us. Haha. Thank you so much for caring about statistics intuition, which is geared toward understanding and thus so much more important than rote memorisation of stats facts that leaves you knowing something but never understanding it.

[@narendeepan]

You literally help me to do all my course work and exam in research methods class in MSc program.

[@lizzyo6028]

Can you explain to me how you do the Huyn-Feldt correction in SPSS?

[@evatsaneva8202]

I would like to ask a question. I still don't get it, and I have read some about covariances, correlation, and the homogeneity assumption..., Why do the groups in the analysis [whether t-test or ANOVA] have to have equal variances? What does this mean in practice? What if they are not equal? Isn't this the aim of the ANOVA analysis (i.e. to find significant differences) ? I really need some help to understand this. I get what is being explained but I don't understand why It is needed and how it translates into practice. Otherwise, thank you very much for the video! It was great :)

[@pagecollector4856]

I dont understand that precisly but I have an idea: If u want to add fractions manually they have to have the same denominators. 2/10 + 3/10 = 5/10. You do 2 + 3 and the 10 stays the same. If they dont have equal denominators it becomes more complicated, 2/10 + 3/12 for example. I think its similar with t-tests and anovas. If the variance is not similar, you are not allowed to do 2 + 3.

[@PedroRibeiro-zs5go]

Thanks, this was helpful!

[@Ganjil100]

Thank you very much! I really appreciate all of your videos. They're super helpful. ^_^

[@johnbaranowski429]

You are very kind to reply to my post. If we are comparing two independent sample means to see if we should reject or fail to reject the null then if the variances are not equal there is a greater risk for type I error??? I was close to understanding with your illustration of two superimposed normal distributions having the same mean but different variance.

[@how2stats]

Correct, if the sample sizes between the two groups are equal.

[@KevinSmith-qt4hz]

@5:04 and still waiting for the "intuition" LOL

[@johnbaranowski429]

Still not understanding sphericity or homogeneity of variance. Why is it important that variances are similar between groups?

[@how2stats]

First, sphericity pertains to repeated measures designs only. It comes down to the fact that statistical analyses like t-test and ANOVA need to estimate a standard error of the difference between the means. In order to estimate the standard error of the difference between the means in a way that keeps alpha near .05, the standard deviations (or variances, more precisely) need to be the same value (within sampling variations) across all groups/conditions, as the method of performing the typical standard error of the mean calculation is to simply take the average of the the variances across the groups/conditions. Averaging the variances presumes they are essentially the same. If they are not the same, then which one should you use to calculate the standard error? That's the problem.

[@evatsaneva8202]

That is exactly what I did not understand as well!

[@Echowable]

Thank you for the video!

[@bosra11]

Explained simply??

[@alijaveed9350]

very clear, thank you

[@Ariel94]

Thank you, this explained it perfectly x

[@tha5289]

Why not just make 1 video instead of dividing them?

[@riveralonzo]

👍

[@OneBlueFroggy]

If this is simplified . . . I'm outa here !