You have a solution for everything Steve! Thank you so much for all your wonderful help!! Caroline ... View more

Hi again Steve, I found a correlation that seems to be positive, but how to get the p-values under each corr coefficient in order to determine if they are significant? Is there a statement that I can include in my code? Proc glimmix data=one; class nem blk year eggtype; model Eggs= eggtype|nem|year/dist=lognormal ddfm=kr; random intercept/subject=blk; random year/residual subject=blk*nem type=ar(1) group=eggtype vcorr; run; Thank you dear Steve! Caroline ... View more

I see. Anyhow, when I plot the data, it doesn´t look like there is really a neg corr with log and logit transformed counts and proportions, eventhough the oscillations seems to be opposite Thank you so much for all your great help Steve! Caroline ... View more

Steve, I was doing a mistake, because I was using the residual option with the beta and the binomial dist, and for distributions where the mean and variance are functionally related you should  not use the residual option.  So now I get negative values with the binomial dist for the prop of healthy vs prop diseased eggs and Gener. Chi-Square / DF=0.08 1 1.0000 -337E-24 9.25E-26 -148E-31 5.01E-33 2 1.0000 -84E-22 2.3E-24 -369E-30 1.25E-31 3 -337E-24 1.0000 -346E-24 5.55E-26 -187E-31 4 -84E-22 1.0000 -863E-23 1.38E-24 -467E-30 5 9.25E-26 -346E-24 1.0000 -271E-24 9.18E-26 6 2.3E-24 -863E-23 1.0000 -676E-23 2.29E-24 7 -148E-31 5.55E-26 -271E-24 1.0000 -262E-24 8 -369E-30 1.38E-24 -676E-23 1.0000 -653E-23 9 5.01E-33 -187E-31 9.18E-26 -262E-24 1.0000 10 1.25E-31 -467E-30 2.29E-24 -653E-23 1.0000 11 1.0000 -301E-24 6.43E-26 -162E-31 4.12E-33 12 1.0000 -749E-23 1.6E-24 -404E-30 1.03E-31 13 -301E-24 1.0000 -342E-24 8.62E-26 -219E-31 14 -749E-23 1.0000 -852E-23 2.15E-24 -546E-30 15 6.43E-26 -342E-24 1.0000 -329E-24 8.37E-26 16 1.6E-24 -852E-23 1.0000 -82E-22 2.08E-24 17 -162E-31 8.62E-26 -329E-24 1.0000 -376E-24 18 -404E-30 2.15E-24 -82E-22 1.0000 -936E-23 19 4.12E-33 -219E-31 8.37E-26 -376E-24 1.0000 20 1.03E-31 -546E-30 2.08E-24 -936E-23 1.0000 Although there is no significant interaction including eggtype in the table for fixed effects. Thanks Steve!! ... View more

I did not get negative values for prop of healthy vs diseased either   Do you think I could use the results (neg corr) from the analysis including number of total eggs (log transformed) and prop of diseased (without log trans) in the same model with gaussian dist, or was this analysis incorrect ? (because you should not compare counts with proportions) Thank you Steve ... View more

Oh noooo!!! now is the problem solved, but I only get positive corr values (I used log without any base) 1 1.0000 0.6982 0.4875 0.3404 0.2377 2 1.0000 0.1012 0.01025 0.001037 0.000105 3 0.6982 1.0000 0.6982 0.4875 0.3404 4 0.1012 1.0000 0.1012 0.01025 0.001037 5 0.4875 0.6982 1.0000 0.6982 0.4875 6 0.01025 0.1012 1.0000 0.1012 0.01025 7 0.3404 0.4875 0.6982 1.0000 0.6982 8 0.001037 0.01025 0.1012 1.0000 0.1012 9 0.2377 0.3404 0.4875 0.6982 1.0000 10 0.000105 0.001037 0.01025 0.1012 1.0000 11 1.0000 0.6982 0.4875 0.3404 0.2377 12 1.0000 0.1012 0.01025 0.001037 0.000105 13 0.6982 1.0000 0.6982 0.4875 0.3404 14 0.1012 1.0000 0.1012 0.01025 0.001037 15 0.4875 0.6982 1.0000 0.6982 0.4875 16 0.01025 0.1012 1.0000 0.1012 0.01025 17 0.3404 0.4875 0.6982 1.0000 0.6982 18 0.001037 0.01025 0.1012 1.0000 0.1012 19 0.2377 0.3404 0.4875 0.6982 1.0000 20 0.000105 0.001037 0.01025 0.1012 1.0000 the logit transformation for the proportions yielded some negative values. Thank you Steve! ... View more

Thanks again Steve! I tried transforming the number of total eggs and he proportion of diseased eggs with log10, and dist=gaussian but did not work ERROR: QUANEW Optimization cannot be completed Optimization routine cannot improve the function value ... View more

Now I compared the proportion of diseased eggs with the number of total eggs (log10 transformed), using dist=gaussian (you did not mean using gaussian and binomial at the same time right?) and it worked!!!! I finally get negative values and the values in the table goes in the right direction (I guess) 1 1.0000 -0.03965 0.001573 -0.00006 2.473E-6 2 1.0000 0.6982 0.4875 0.3404 0.2377 3 -0.03965 1.0000 -0.03965 0.001573 -0.00006 4 0.6982 1.0000 0.6982 0.4875 0.3404 5 0.001573 -0.03965 1.0000 -0.03965 0.001573 6 0.4875 0.6982 1.0000 0.6982 0.4875 7 -0.00006 0.001573 -0.03965 1.0000 -0.03965 8 0.3404 0.4875 0.6982 1.0000 0.6982 9 2.473E-6 -0.00006 0.001573 -0.03965 1.0000 10 0.2377 0.3404 0.4875 0.6982 1.0000 11 1.0000 -0.03965 0.001573 -0.00006 2.473E-6 12 1.0000 0.6982 0.4875 0.3404 0.2377 13 -0.03965 1.0000 -0.03965 0.001573 -0.00006 14 0.6982 1.0000 0.6982 0.4875 0.3404 15 0.001573 -0.03965 1.0000 -0.03965 0.001573 16 0.4875 0.6982 1.0000 0.6982 0.4875 17 -0.00006 0.001573 -0.03965 1.0000 -0.03965 18 0.3404 0.4875 0.6982 1.0000 0.6982 19 2.473E-6 -0.00006 0.001573 -0.03965 1.0000 20 0.2377 0.3404 0.4875 0.6982 1.0000 ... View more