This is my program and covariance parameter estimates.  Is there a similar estimate?   proc mixed data=Result; class subject period treatment time; model result = baseline period subject treatment time treatment*time; repeated time / type=CS subject=subject*period lsmeans treatment / adjust=dunnett;   Covariance parameter estimates Cov Par    Subject                 Estimate CS             subject*period    3.7383 Residual                                 43.5428 ... View more

We're mostly concerned with the pairwise comparisons following a significant treatment or treatment by time interaction, at which point each timepoint is analyzed with the following code:   proc mixed data=Result; class subject treatment; model result = baseline subject treatment; lsmeans treatment / adjust=dunnett; run; quit; Finding the standard error of the difference in lsmeans is easy enough, but I'm not sure how to find the 2 t-values [t(.05,ddf) + t(.80,ddf)], which will be multiplied by the stderr of the difference in lsmeans.  Here is an example output: Thanks!   ... View more

Correct, both of those models with proc mixed (just using a different data set). I’m curious as to how sas treats the data differently using the first vs 2nd model you mention. I’m curious because it’s still showing a valid significant treatmentxtime interaction as it should, and the results are not far off when I add dunnetts test to compare lsmeans between treatments. So it seems to still be making valid inferences…I just not sure how it is treating the data at repeated timepoints without setting a subject= option in the repeated statement. ... View more

It is repeated measures data at four timepoints 1-4 hours. Without using a subject=option my data is giving me the correct significance for treatment and treatment*time interaction…except the pvalues are lower than when a subject= option is entered. So I’m curious to how it is treating the data in the four timepoints without defining subject, especially since I’m getting results similar to what is given with a subject= option in the repeated line. ... View more

Thanks! If the subject= statement is not included and the data is not correlated, is this still considered a “repeated measures” analysis? The results are giving pvalues for the treatment*time interaction that are indicating the correct trend is being detected, but I wasn’t sure if there was real utility in doing such an analysis (picking up the same trends, with less power?). Thanks! ... View more

Thanks! My last question is when I remove the subject= statement completely but leave in the repeated time, the results are very similar. Can you explain why the results are similar when it is removed and how the analysis methods are different this way? It actually has better sensitivity to pick up changes, but wouldn’t use it if it was wrong. ... View more

Would it be appropriate to include group=treatment in this case? Thanks! ... View more

With the following code I'm receiving a LinTRT*time p-value of 0.1542 (should be 0.105 based on my reference).  Everything under the "Type 3 Tests of Fixed Effects" is accurate and so is the Linear treatment effect from the -3 -1 1 3 contrast.  Any ideas on what would be causing the p-value discrepancy?   data Result; input treatment time subject period result baseline; datalines; 1 1 1 1 112 109 1 2 1 1 113 109 1 3 1 1 105 109 1 4 1 1 111 109 1 5 1 1 102 109 1 6 1 1 111 109 1 1 2 8 102 105 1 2 2 8 103 105 1 3 2 8 106 105 1 4 2 8 104 105 1 5 2 8 107 105 1 6 2 8 101 105 1 1 3 15 98 99 1 2 3 15 95 99 1 3 3 15 94 99 1 4 3 15 96 99 1 5 3 15 92 99 1 6 3 15 94 99 1 1 4 22 102 106 1 2 4 22 109 106 1 3 4 22 105 106 1 4 4 22 113 106 1 5 4 22 113 106 1 6 4 22 111 106 2 1 1 22 112 112 2 2 1 22 112 112 2 3 1 22 109 112 2 4 1 22 113 112 2 5 1 22 109 112 2 6 1 22 113 112 2 1 2 1 105 101 2 2 2 1 108 101 2 3 2 1 99 101 2 4 2 1 103 101 2 5 2 1 98 101 2 6 2 1 104 101 2 1 3 8 96 96 2 2 3 8 96 96 2 3 3 8 96 96 2 4 3 8 99 96 2 5 3 8 99 96 2 6 3 8 94 96 2 1 4 15 101 106 2 2 4 15 108 106 2 3 4 15 111 106 2 4 4 15 114 106 2 5 4 15 105 106 2 6 4 15 110 106 3 1 1 15 118 114 3 2 1 15 111 114 3 3 1 15 113 114 3 4 1 15 111 114 3 5 1 15 110 114 3 6 1 15 113 114 3 1 2 22 105 112 3 2 2 22 115 112 3 3 2 22 110 112 3 4 2 22 117 112 3 5 2 22 112 112 3 6 2 22 108 112 3 1 3 1 95 92 3 2 3 1 93 92 3 3 3 1 95 92 3 4 3 1 100 92 3 5 3 1 93 92 3 6 3 1 93 92 3 1 4 8 107 109 3 2 4 8 115 109 3 3 4 8 113 109 3 4 4 8 112 109 3 5 4 8 113 109 3 6 4 8 105 109 4 1 1 8 117 112 4 2 1 8 106 112 4 3 1 8 111 112 4 4 1 8 109 112 4 5 1 8 108 112 4 6 1 8 106 112 4 1 2 15 112 113 4 2 2 15 110 113 4 3 2 15 112 113 4 4 2 15 109 113 4 5 2 15 108 113 4 6 2 15 109 113 4 1 3 22 104 96 4 2 3 22 92 96 4 3 3 22 95 96 4 4 3 22 92 96 4 5 3 22 98 96 4 6 3 22 92 96 4 1 4 1 109 107 4 2 4 1 110 107 4 3 4 1 107 107 4 4 4 1 109 107 4 5 4 1 106 107 4 6 4 1 108 107 ; Run; proc mixed data=Result; class subject period treatment time; model result = baseline period subject treatment time treatment*time; repeated time / type=CS subject=subject*period; lsmeans treatment / pdiff=control; contrast 'linear' treatment -3 -1 1 3; lsmestimate Treatment*time 'Linear treatment at time 1' -3 0 0 0 0 0 -1 0 0 0 0 0 1 0 0 0 0 0 3 0 0 0 0 0, 'Linear treatment at time 2' 0 -3 0 0 0 0 0 -1 0 0 0 0 0 1 0 0 0 0 0 3 0 0 0 0, 'Linear treatment at time 3' 0 0 -3 0 0 0 0 0 -1 0 0 0 0 0 1 0 0 0 0 0 3 0 0 0, 'Linear treatment at time 4' 0 0 0 -3 0 0 0 0 0 -1 0 0 0 0 0 1 0 0 0 0 0 3 0 0, 'Linear treatment at time 5' 0 0 0 0 -3 0 0 0 0 0 -1 0 0 0 0 0 1 0 0 0 0 0 3 0, 'Linear treatment at time 6' 0 0 0 0 0 -3 0 0 0 0 0 -1 0 0 0 0 0 1 0 0 0 0 0 3/FTEST; run; quit; ... View more

I'm currently using the following code and am receiving the correct p-values for the overall treatment effect. can you provide any advice on the coefficients to use to get the correct overall linear treatment*time interaction p-value? I'm not interested in looking at individual timepoints unless the pooled overall Treatment*time effect is significant. This is a crossover with 4 subjects/group, control + 3 treatment groups. 6 timepoints. proc mixed data=Result; class subject period treatment time; model result = treatment time period subject treatment*time; repeated time / type=CS subject=subject*period; lsmeans treatment / pdiff=control; lsmestimate Treatment 3 -1 1 3; lsmestimate Treatment*time <need help with these coefficients>; ... View more

Thanks I will try that.  I did receive some code however I don't understand what the periods mean in the lsmestimate statement ( 3 0 0 0.. -1 0 0 0.. 1 0 0 0.. 3 0 0 0) and how I adjust for my specific design?     proc mixed data=Result; class subject period treatment time; model result = treatment time period subject treatment*time; repeated time / type=CS subject=subject*period; lsmeans treatment / pdiff=control; lsmestimate Treatment*time 3 0 0 0.. -1 0 0 0.. 1 0 0 0.. 3 0 0 0; ... View more

I need to evaluate dose response relationship through linear trend tests.  I first ran the test for linear treatment for the overall treatment effect.  I also need to evaluate the linearTreatment*time.  If linear Treatment*time is significant I will run the tests at each individual timepoint.  But the p-value I'm receiving for the LinearTreatment*time interaction is not matching my reference data so am looking for help with the code. ... View more