I am currently working on some multiple testing adjustments for a range of regression models that I am running on a dataset from a randomised control trial. In order to store my p values into a new datast, I was planning to use the parmest package in Stata offered by Roger Newson.
I have set up my code as suggested here http://www.stata.com/meeting/8uk/transp1.pdf but keep getting the same error notice when trying to concatenate the output files (after dsconcat ‘tf1’ ‘tf2’ ‘tf3’): "No input data sets have been specified". Any advice on how to refine my code?
tempfile tf1 tf2 tf3;
parmby "regress SA_needs_CT3 TrialArm SA_needs_CT1a rural, vce(cluster ClusterID_AT1a)",label idnum(1) idstr(Unadj.) saving(`tf1',replace);
parmby "reg SA_enough_CT3 TrialArm SA_enough_CT1a rural, vce(cluster ClusterID_AT1a)",label idnum(2) idstr(Unadj.) saving(`tf2',replace);
parmby "reg ESS_pca_C3 TrialArm rural ESS_pca_C1, vce(cluster ClusterID_AT1a)",label idnum(3) idstr(Adj.) saving(`tf3',replace);
dsconcat `tf1' `tf2' `tf3';
format estimate min95 max95 %8.2f;
sort idnum idstr parmseq;
by idnum idstr:list parm label estimate min95 max95,noobs;
I have set up my code as suggested here http://www.stata.com/meeting/8uk/transp1.pdf but keep getting the same error notice when trying to concatenate the output files (after dsconcat ‘tf1’ ‘tf2’ ‘tf3’): "No input data sets have been specified". Any advice on how to refine my code?
tempfile tf1 tf2 tf3;
parmby "regress SA_needs_CT3 TrialArm SA_needs_CT1a rural, vce(cluster ClusterID_AT1a)",label idnum(1) idstr(Unadj.) saving(`tf1',replace);
parmby "reg SA_enough_CT3 TrialArm SA_enough_CT1a rural, vce(cluster ClusterID_AT1a)",label idnum(2) idstr(Unadj.) saving(`tf2',replace);
parmby "reg ESS_pca_C3 TrialArm rural ESS_pca_C1, vce(cluster ClusterID_AT1a)",label idnum(3) idstr(Adj.) saving(`tf3',replace);
dsconcat `tf1' `tf2' `tf3';
format estimate min95 max95 %8.2f;
sort idnum idstr parmseq;
by idnum idstr:list parm label estimate min95 max95,noobs;
Comment