Create flagged variable based on multiple grader responses

[Abishekaa]

Hi all,

         Thank you in advance! I have a dataset with multiple (6) grader responses. I need to create flag variables based on these responses:

1- One flagged variable if all responses match

2- One flagged variable if all except one response match

3- One flag for no response matching (not 1 or 2)

I have trouble creating the flag for 2 -- how do I find responses where all except one match?

Here is a sample dataset:

num rater1 rater2 rater3 rater4 rater5 rater6  Flag2
Pt1         4         4        4          4        4      4       0

Pt2         3         4        4          4        4      4       1

Pt3         1         2        3          3        2      1       0

Pt4         1         2         1         1         1     1       1

How do I create the 'Flag2' variable which flags any observations where all responses except one match between raters?

[Reeza · Accepted solution]

Reframing your logic:

• Flag1 = Max of series = Min of series - all are the same
• Flag2 = Max of series is repeated 5 times or Min of series is repeated 5 times. Can switch 5 to DIM(of rater(*))-1 to be more dynamic. 
• Flag3 = Flag1=0 & Flag2 = 0

data have;
input num $ rater1 rater2 rater3 rater4 rater5 rater6  Flag2;
cards;
Pt1         4         4        4          4        4      4       0
Pt2         3         4        4          4        4      4       1
Pt3         1         2        3          3        2      1       0
Pt4         1         2         1         1         1     1       1
;;;;
run;

data want;
set have;
array rater(6) rater1-rater6;

flag1 = max(of rater(*)) = min(of rater(*));

flag_check = flag1=0 and 
(countc(catx(",", of rater(*)), max(of rater(*))) = 5 or
countc(catx(",", of rater(*)), min(of rater(*))) = 5);

flag3 = flag1=0 & flag2=0;

run;

Here's the more traditional version using IF/THEN. 

data want;
set have;
array rater(6) rater1-rater6;

if max(of rater(*)) = min(of rater(*)) then flag1=1; else flag1=0;

if flag1=0 and 
(countc(catx(",", of rater(*)), max(of rater(*))) = (Dim(rater)-1) or
countc(catx(",", of rater(*)), min(of rater(*))) = (dim(rater)-1)) then flag2=1;
else flag2=0;

if flag1=0 and flag2=0 then flag3 = 1;
else flag3=0;



run;

---

@Abishekaa wrote:

Hi all,

         Thank you in advance! I have a dataset with multiple (6) grader responses. I need to create flag variables based on these responses:

1- One flagged variable if all responses match

2- One flagged variable if all except one response match

3- One flag for no response matching (not 1 or 2)

I have trouble creating the flag for 2 -- how do I find responses where all except one match?

Here is a sample dataset:

num rater1 rater2 rater3 rater4 rater5 rater6  Flag2
Pt1         4         4        4          4        4      4       0

Pt2         3         4        4          4        4      4       1

Pt3         1         2        3          3        2      1       0

Pt4         1         2         1         1         1     1       1

How do I create the 'Flag2' variable which flags any observations where all responses except one match between raters?

---

[Abishekaa]

Hi Reeza,

                 Thanks so much! Could you be kind enough to explain how this flag could be created if the data are character variables instead (example below)?

num rater1 rater2 rater3 rater4 rater5 rater6  Flag2
Pt1         a         a        a          a        a      a       0

Pt2         b         a        a          a        a      a       1

Pt3         a         b        c          c        b      a       0

Pt4         c         a         c          c        c      c       1

[Ksharp]

data have;
input (num rater1 rater2 rater3 rater4 rater5 rater6) ($) ;
cards;
Pt1         a         a        a          a        a      a       
Pt2         b         a        a          a        a      a       
Pt3         a         b        c          c        b      a       
Pt4         c         a         c          c        c      c  
;
run;
data want;
if _n_=1 then do;
 length k $ 80;
 declare hash h();
 h.definekey('k');
 h.definedone();
end;
set have;
array x{*} $ rater1-rater6;
do i=1 to dim(x);
 if not missing(x{i}) then do;k=x{i}; h.ref();end;
end;
flag=h.num_items=2;
h.clear();
drop i k;
run;

proc print;run;