# Which test to use?

#### BALA M

##### New Member
I have 3 groups-A, B, C.
No. of respondents in each group A-120, B-120, C-30.
Each group has an index about a topic. There are 14 such topics.

In short, 14 indices of A, B and C.
What is the statistical test to find if there is significant difference between the indices of the 3 groups?

I did one way ANOVA and it said no significant difference.
I got confused :shakehead because the maximum value of the index being 100 on any topic, one topic gave indices like this 60, 86,88 by A, B & C respectively.

Isn't that a big difference?
Also what test to do to find the concordance between each group?

#### Karabiner

##### TS Contributor
If there wasn't a mistake in the calculation, then the non-significant
result could maybe be due to huge variance. That is, a mean difference
of 60 vs. 88 is large if the standard deviation is 10, for example, but
the same difference is much smaller if standard deviation was 100.
Maybe there are outliers? Moreover, if your smallest group also has
a markedly smaller variance than one or two of the other groups, this
might result in reduced statistical power of the test.

With kind regards

K.