# Chi-square test - What do I do next?

B

#### big_tj

Hey guys

I did a Chi-square test and found out that that there was a statiscally significant difference among figures from my 3 units. I was testing to see if there was a statistical difference between the % (yes/no) of some results of 3 units.

Now that there is a difference that is statiscally signifant, what tool do I use to find out which is statistically differnent from the others? I was thinking that i might have to do different chi-square tests on the result of each unit two at a time and see which "set" produces a statistically significant variance; would that work? If this approach does work, is there a simplier way - I ultimately need to do some comparison between results of approximately 10 units.

Also, how can i statistically show which result is the top result? Thanks!

Last edited by a moderator:

#### Steve Prevette

##### Deming Disciple
Staff member
Super Moderator
What I do is look at the chi square inputs from each of the individuals. Any that exceed 9 (three standard deviations) are likely significantly different from the others, any that exceed 4 (two standard deviations) are suspect.

input = (Measured - Expected)squared divided by Expected

B

#### big_tj

What I do is look at the chi square inputs from each of the individuals. Any that exceed 9 (three standard deviations) are likely significantly different from the others, any that exceed 4 (two standard deviations) are suspect.

input = (Measured - Expected)squared divided by Expected
Hmmm...interesting perspective; never thought of that. Thanks a lot; I will see if that works!

Can you expound a bit more on the 9 equating to three standard deviation?

Last edited by a moderator: