/* this program undertakes renewed analyses following rejected paper on syllable structure submitted in Spring 08. */


libname sli "C:\1dataLap\MASHA_SLI\";
run;

PROC IMPORT OUT= sli.realnonsens1 
            DATAFILE= "C:\1dataLap\MASHA_SLI\SLIForStatsAll090908.xls" 
            DBMS=EXCEL REPLACE;
     
     GETNAMES=YES;
     MIXED=NO;
     SCANTEXT=YES;
     USEDATE=YES;
     SCANTIME=YES;
RUN;

proc contents data = sli.realnonsens1; run;
proc print; var F13-F18; run;

proc freq data = sli.realnonsens1; tables subject; run;
proc sort data = sli.realnonsens1; by subject; run;
proc print data = sli.realnonsens1; var subject corrections MLU_Match Pseudo_Rom; run;


proc freq; tables mlu_match age_match type; run;

proc format;
value $resp '1' = 'correct' '0' = 'wrong'; 
run;
proc print; var corrections; run;

data x; set sli.realnonsens1;
 if subject = 1 or subject = 2 or subject = 3 or subject = 4 or subject = 5 or subject = 6  or subject = 7 then group_chr = 'SLI';
 if subject = 8 or subject = 9 or subject = 10 or subject = 11 or subject = 12 or subject = 13 or subject = 14 then group_chr = 'MLU';
 if subject = 15 or subject = 16 or subject = 17 or subject = 18 or subject = 19 or subject = 20 or subject = 21 then group_chr = 'Age'; 

 syl_struct = '           ';
 if corrections = 'CCV_1_1' then syl_struct = 'CCV';
 if corrections = 'CCV_1_2' then syl_struct = 'CCV';
 if corrections = 'CCV_2_1' then syl_struct = 'CCV';
 if corrections = 'CCV_2_2' then syl_struct = 'CCV';
 if corrections = 'CCV_3_1' then syl_struct = 'CCV';
 if corrections = 'CCV_3_2' then syl_struct = 'CCV';

if corrections = 'CCVC_1_1' then syl_struct = 'CCVC';
if corrections = 'CCVC_1_2' then syl_struct = 'CCVC';
if corrections = 'CCVC_2_1' then syl_struct = 'CCVC';
if corrections = 'CCVC_2_2' then syl_struct = 'CCVC';
if corrections = 'CCVC_3_1' then syl_struct = 'CCVC';
if corrections = 'CCVC_3_2' then syl_struct = 'CCVC';
if corrections = 'CCVC_3_3' then syl_struct = 'CCVC';


if corrections = 'CCVCC_1_1' then syl_struct = 'CCVCC';
if corrections = 'CCVCC_1_2' then syl_struct = 'CCVCC';
if corrections = 'CCVCC_2_1' then syl_struct = 'CCVCC';
if corrections = 'CCVCC_2_2' then syl_struct = 'CCVCC';
if corrections = 'CCVCC_3_1' then syl_struct = 'CCVCC';
if corrections = 'CCVCC_3_2' then syl_struct = 'CCVCC';
if corrections = 'CCVCC_3_3' then syl_struct = 'CCVCC';

if corrections = 'CV_1_1' then syl_struct = 'CV';
if corrections = 'CV_1_2' then syl_struct = 'CV';
if corrections = 'CV_2_1' then syl_struct = 'CV';
if corrections = 'CV_2_2' then syl_struct = 'CV';
if corrections = 'CV_3_1' then syl_struct = 'CV';
if corrections = 'CV_3_2' then syl_struct = 'CV';
if corrections = 'CV_3_3' then syl_struct = 'CV';

if corrections = 'CVC_1_1' then syl_struct = 'CVC';
if corrections = 'CVC_1_2' then syl_struct = 'CVC';
if corrections = 'CVC_2_1' then syl_struct = 'CVC';
if corrections = 'CVC_2_2' then syl_struct = 'CVC';
if corrections = 'CVC_3_1' then syl_struct = 'CVC';
if corrections = 'CVC_3_2' then syl_struct = 'CVC';
if corrections = 'CVC_3_3' then syl_struct = 'CVC';

if corrections = 'CVCC_1_1' then syl_struct = 'CVCC';
if corrections = 'CVCC_1_2' then syl_struct = 'CVCC';
if corrections = 'CVCC_2_1' then syl_struct = 'CVCC';
if corrections = 'CVCC_2_2' then syl_struct = 'CVCC';
if corrections = 'CVCC_3_1' then syl_struct = 'CVCC';
if corrections = 'CVCC_3_2' then syl_struct = 'CVCC';
if corrections = 'CVCC_3_3' then syl_struct = 'CVCC';

if corrections = 'CCV_1_1' then syl_struct = 'CCV';
if corrections = 'CCV_1_2' then syl_struct = 'CCV';
if corrections = 'CCV_2_1' then syl_struct = 'CCV';
if corrections = 'CCV_2_2' then syl_struct = 'CCV';
if corrections = 'CCV_3_1' then syl_struct = 'CCV';
if corrections = 'CCV_3_2' then syl_struct = 'CCV';
if corrections = 'CCV_3_3' then syl_struct = 'CCV';

if corrections = 'VCC_1_1' then syl_struct = 'VCC';
if corrections = 'VCC_1_2' then syl_struct = 'VCC';
if corrections = 'VCC_2_1' then syl_struct = 'VCC';
if corrections = 'VCC_2_2' then syl_struct = 'VCC';
if corrections = 'VCC_3_1' then syl_struct = 'VCC';
if corrections = 'VCC_3_2' then syl_struct = 'VCC';
if corrections = 'VCC_3_3' then syl_struct = 'VCC';

if corrections = 'VC_1_1' then syl_struct = 'VC';
if corrections = 'VC_1_2' then syl_struct = 'VC';
if corrections = 'VC_2_1' then syl_struct = 'VC';
if corrections = 'VC_2_2' then syl_struct = 'VC';
if corrections = 'VC_3_1' then syl_struct = 'VC';
if corrections = 'VC_3_2' then syl_struct = 'VC';
if corrections = 'VC_3_3' then syl_struct = 'VC';


 syl_no = '           ';
 if corrections = 'CCV_1_1' then syl_no = 'One';
 if corrections = 'CCV_1_2' then syl_no = 'One';
 if corrections = 'CCV_2_1' then syl_no = 'Two';
 if corrections = 'CCV_2_2' then syl_no = 'Two';
 if corrections = 'CCV_3_1' then syl_no = 'Three';
 if corrections = 'CCV_3_2' then syl_no = 'Three';

if corrections = 'CCVC_1_1' then syl_no = 'One';
if corrections = 'CCVC_1_2' then syl_no = 'One';
if corrections = 'CCVC_2_1' then syl_no = 'Two';
if corrections = 'CCVC_2_2' then syl_no = 'Two';
if corrections = 'CCVC_3_1' then syl_no = 'Three';
if corrections = 'CCVC_3_2' then syl_no = 'Three';
if corrections = 'CCVC_3_3' then syl_no = 'Three';


if corrections = 'CCVCC_1_1' then syl_no = 'One';
if corrections = 'CCVCC_1_2' then syl_no = 'One';
if corrections = 'CCVCC_2_1' then syl_no = 'Two';
if corrections = 'CCVCC_2_2' then syl_no = 'Two';
if corrections = 'CCVCC_3_1' then syl_no = 'Three';
if corrections = 'CCVCC_3_2' then syl_no = 'Three';
if corrections = 'CCVCC_3_3' then syl_no = 'Three';

if corrections = 'CV_1_1' then syl_no = 'One';
if corrections = 'CV_1_2' then syl_no = 'One';
if corrections = 'CV_2_1' then syl_no = 'Two';
if corrections = 'CV_2_2' then syl_no = 'Two';
if corrections = 'CV_3_1' then syl_no = 'Three';
if corrections = 'CV_3_2' then syl_no = 'Three';
if corrections = 'CV_3_3' then syl_no = 'Three';

if corrections = 'CVC_1_1' then syl_no = 'One';
if corrections = 'CVC_1_2' then syl_no = 'One';
if corrections = 'CVC_2_1' then syl_no = 'Two';
if corrections = 'CVC_2_2' then syl_no = 'Two';
if corrections = 'CVC_3_1' then syl_no = 'Three';
if corrections = 'CVC_3_2' then syl_no = 'Three';
if corrections = 'CVC_3_3' then syl_no = 'Three';

if corrections = 'CVCC_1_1' then syl_no = 'One';
if corrections = 'CVCC_1_2' then syl_no = 'One';
if corrections = 'CVCC_2_1' then syl_no = 'Two';
if corrections = 'CVCC_2_2' then syl_no = 'Two';
if corrections = 'CVCC_3_1' then syl_no = 'Three';
if corrections = 'CVCC_3_2' then syl_no = 'Three';
if corrections = 'CVCC_3_3' then syl_no = 'Three';

if corrections = 'CCV_1_1' then syl_no = 'One';
if corrections = 'CCV_1_2' then syl_no = 'One';
if corrections = 'CCV_2_1' then syl_no = 'Two';
if corrections = 'CCV_2_2' then syl_no = 'Two';
if corrections = 'CCV_3_1' then syl_no = 'Three';
if corrections = 'CCV_3_2' then syl_no = 'Three';
if corrections = 'CCV_3_3' then syl_no = 'Three';

if corrections = 'VCC_1_1' then syl_no = 'One';
if corrections = 'VCC_1_2' then syl_no = 'One';
if corrections = 'VCC_2_1' then syl_no = 'Two';
if corrections = 'VCC_2_2' then syl_no = 'Two';
if corrections = 'VCC_3_1' then syl_no = 'Three';
if corrections = 'VCC_3_2' then syl_no = 'Three';
if corrections = 'VCC_3_3' then syl_no = 'Three';

if corrections = 'VC_1_1' then syl_no = 'One';
if corrections = 'VC_1_2' then syl_no = 'One';
if corrections = 'VC_2_1' then syl_no = 'Two';
if corrections = 'VC_2_2' then syl_no = 'Two';
if corrections = 'VC_3_1' then syl_no = 'Three';
if corrections = 'VC_3_2' then syl_no = 'Three';
if corrections = 'VC_3_3' then syl_no = 'Three';

Group = '             ';
If type = 1 then Group = 'SLI';
If type = 2 then Group = 'MLU_match';
If type = 3 then Group = 'Age_match';

Corr_resp1 = .; 
If corr_resp = 0 then Corr_resp1 = 1;
If corr_resp = 1 then Corr_resp1 = 0;

/*numeric to character conversion example for class*/
Corr_resp_chr = put(Corr_resp1, 1.);
/*character to numeric conversion example for class*/
Corr_resp2 = input(Corr_resp_chr, 8.);
*if Corr_resp_chr = 1 then Corr_resp_chr = 'Correct';
*if Corr_resp_chr = 0 then Corr_resp_chr = 'Wrong';
format Corr_resp_chr $resp.; 
drop F13-F18;
run;

proc print data = x; var Corr_resp_chr corr_resp Corr_resp1 Corr_resp2; run;
proc contents; run;

proc means data = x; title 'Means by subject across sylno and structure'; by subject; var corr_resp1; run;

proc freq data = x; tables Corr_resp_chr Group corr_resp_chr * Group/chisq; run;
proc print data = x; var syl_struct syl_no corrections group; run;
proc print data = x; var initial_cluster medial_cluster final_cluster; run;
proc freq data = x; tables initial_cluster medial_cluster final_cluster; run;
Proc freq data = x; tables Group*group_chr; run;
proc print data = x; where corr_resp = .; var Subject pseudo_rom Corr_resp; run;
proc sort data = x; by group_chr; run;
proc means data = x; Title 'Means by Group and Syllable Number'; by group_chr;
   var corr_resp;
   *output out=means_sylno mean=Repetitions;
run;

proc freq data = x; title 'Number of Correct Responses by Group'; where Group_chr ne 'Age'; tables Group_chr*corr_resp /chisq; run;
proc freq data = x; title 'Number of Correct Responses by Group'; where Group_chr ne 'MLU'; tables Group_chr*corr_resp /chisq; run;
proc freq data = x; title 'Number of Correct Responses by Group and Subject';  tables Subject*Group_chr*corr_resp; run;
Proc freq data = x; title 'Number of Correct Responses by Nonsense Word'; tables pseudo_rom*corr_resp; run;
Proc freq data = x; title 'Number of Correct Responses by Group and Nonsense Word'; tables Group_chr*pseudo_rom*corr_resp; run;

proc freq data = x; tables corr_resp1 syl_no; run;

proc sort data = x;
by GROUP syl_no syl_struct;
run;


/*very possibly a statistically inappropriate analysis for this binary outcome, but one that the field uses 
for this particular situations*/
proc glm data = x;
Title 'Syl_number and structure with Group';
 *where group ne ' ';
      class GROUP_chr syl_no syl_struct; 
      model Corr_resp1 = GROUP_chr|syl_no|syl_struct; 
      lsmeans GROUP_chr|syl_no|syl_struct /pdiff out=lsm;
	  means GROUP_chr|syl_no|syl_struct ;
run; 
quit;

/*experimenting with Proc Genmod for binary outcome*/
/*RUN THE LOGISTIC REGRESSION USING PROC GENMOD*/
title "Logistic Regression Using Proc Genmod";
proc genmod data=x descending;
 class Group_chr Syl_no Syl_struct / param=ref;
  model Corr_resp1 = Group_chr  syl_no syl_struct /dist=bin type3;
run;

PROC GENMOD DATA=x descending ;
CLASS Subject Group_chr Syl_no Syl_struct ;
MODEL Corr_resp1 = Group_chr Syl_no Syl_struct Group_chr*Syl_no Syl_no*Syl_struct 
                / dist=binomial link=logit type3 ;
REPEATED subject=subject / type=ar(1) CORRW;
LSMEANS Group_chr / diff cl ;
RUN;


/*try above with logit link.  */

/*some charting*/

proc chart data=x;
   vbar corr_resp / group=group_chr; run;
ref=100 200 300
sumvar=corr_resp
type=mean;
run;