Testing Phenotypic Ratios: Results of Chi Square Test

[jena]

Hi,
My Question:
Problem 1. The expected phenotypic ratios for each of the crosses below is 1:1. Determine whether the observed ratios are differ significantly (p < 0.05) from the expected.
View attachment Table 1.doc
I used a Chi Square test to help me determine that answer and found the there is no probability of any of them correct.
Answers:
•Let’s first examine white eyed females to red-eyed males

o We found that X2: Results
o 2 data/expectation pairs (x,E):
( 225. , 433.0 ); ( 208. , 433.0 );
o chi-square = 217.
degrees of freedom = 1
probability = 0.000​

• Next we examined Red-eyed females to eosin eyed males

o We found that X2: Results
o 2 data/expectation pairs (x,E):
( 679. , 1426. ); ( 747. , 1426. );
o chi-square = 715. degrees of freedom = 1 probability = 0.000​

• Finally we examined Eosin-eyed females to white eyed females

o We found that X2: Results
o 2 data/expectation pairs (x,E):
( 694. , 1273. ); ( 579. , 1273. );
o chi-square = 642.
degrees of freedom = 1
probability = 0.000​

Does answer seem correct?
Thank You

 

[Moonbear]

•Let’s first examine white eyed females to red-eyed males

o We found that X2: Results
o 2 data/expectation pairs (x,E):
( 225. , 433.0 ); ( 208. , 433.0 );
o chi-square = 217.
degrees of freedom = 1
probability = 0.000​

I think we can just go through one of these, and then you can follow the same concept for all of them. You want to compare your data (actual or observed values) to the expected values for a given trait. Your total number of observations (n) is 433, but that's not your expected value. If you are predicting a 1:1 ratio, that means you're expecting 50% to be of each trait (if ratios don't make a lot of sense to you, deal a deck of cards into two piles...for every 1 you put in pile A, put 1 in pile B...what percentage of the deck of cards is in pile A?). So, your expected value for either trait is going to be 50% of the total number of observations. That should change your outcome substantially if you redo your Chi-squared calculations with that information.