# Thread: Need clarifications with X^2 (chi squared) data - Punnet Squares

1. ## Need clarifications with X^2 (chi squared) data - Punnet Squares

I have the following data for fruit flies :

Mutant male : 23
normal male : 28

Mutant female : 44
normal female : 41

So i made a punnet square and then figured that 1/2 the population should be male. 1/2 the male population should be mutants. And half the female population should be mutants.

Obviously its not 50-50 on male-female HOWEVER, when i calculated my E value, instead of multiplying 25% with the total population, i multipled 50% to the total female/male population respectively.

So the expected values for male would be .50* (23+28) = 25.5

Square that, and divide by e, and i get my chi squared values. I do this for all the values and find my null hypothesis is correct.

HOWEVER

at school, my teacher said that was incorrect (he couldn't explain why), that we had to multiply the total population by 25% , and not 50% of the the male/female pop.

Why is that? Like, 50% of 50% is still 25%, and this way it allows my data to fall within reasonable limits. I just want a clarification on why i have to do 25% of the total population and cannot do 50% of the male/female populations for my "e"

2. ## Re: Need clarifications with X^2 (chi squared) data - Punnet Squares

Hey Tideo.

In this example you have four probabilities so your distrribution will have four values instead of two.

From this distribution you can get a conditional distribution conditioning on any variable and then using that you can take expectations or get variances.

Consider looking at the whole picture and then condition on that.