the aim is to decide whether gene types can affect response to certain stimulus.

there are 3 types of genes(G1-3), 3 kinds of responses(R1-3). in one experiment, dozens individuals from each gene group are involved, some of them (a) showed response 1, some showed R2, etc.

so the contingency table is like this:

R1 R2 R3

G1 a b c

G2 d e f

G3 g h i

but the problem is that this experiment was repeated for 10 times and each time we got different data.

we are asked to draw contingency table based on the whole data and do Chi-square test.

what ive tried:

i summed up all the frequency from 10 experiments and get a contingency table:

R1 R2 R3

G1 a1-10 b 1-10 c1-10

G2 d1-10 e1-20 f1-10

G3 g1-10 h 1-10 i1-10

did the Chi test and P<0.05

i also tried to use the [average] (instead of [sum])to represent the whole: and this time: P>0.05

i cannt use McNemar since its not 2x2.

obviously, neither of the solution is reliable.

so, can i get some ideas?

or, should i try another statistic method like ANOVA(no idea how to fit this data in ANOVA) or logistic regressoin (no idea)?

Thanks.

there are 3 types of genes(G1-3), 3 kinds of responses(R1-3). in one experiment, dozens individuals from each gene group are involved, some of them (a) showed response 1, some showed R2, etc.

so the contingency table is like this:

R1 R2 R3

G1 a b c

G2 d e f

G3 g h i

but the problem is that this experiment was repeated for 10 times and each time we got different data.

we are asked to draw contingency table based on the whole data and do Chi-square test.

what ive tried:

i summed up all the frequency from 10 experiments and get a contingency table:

R1 R2 R3

G1 a1-10 b 1-10 c1-10

G2 d1-10 e1-20 f1-10

G3 g1-10 h 1-10 i1-10

did the Chi test and P<0.05

i also tried to use the [average] (instead of [sum])to represent the whole: and this time: P>0.05

i cannt use McNemar since its not 2x2.

obviously, neither of the solution is reliable.

so, can i get some ideas?

or, should i try another statistic method like ANOVA(no idea how to fit this data in ANOVA) or logistic regressoin (no idea)?

Thanks.

Last edited: