01-09-2017 01:14 PM
I import a data set from SPSS but the responses are displayed as texts. I want them to be recoded into values that correspond to Strongly Agree = 7, Agree = 6.................. and Strongly Disagree = 1
01-09-2017 01:45 PM
Please post your code as TEXT not as an image, otherwise we need to type out your code to make suggestions.
Text needs to be quoted when used in comparisons, otherwise it's looking for a variable. Quote the characters in your recode statement.
01-09-2017 02:03 PM
You cannot change the variable type from character to numeric for an existing SAS variable. Best, for a number of reasons would be recode into a new varaible that would be numeric. I prefer to recode using a custom informat for things like this, especially when there are many variables. Here is a brief example creating an informat and manipulating some dummy data.
proc format library=work; invalue Likert (upcase) "AGREE"=6 "AGREE SOMEWHAT"=5 "NEUTRAL" = 4 "DISAGREE SOMEWHAT"=3 "DISAGREE" = 2 ; RUN; /* creating some example data*/ Data have; infile datalines dlm=','; informat Q10_1 Q10_2 $20.; input Q10_1 Q10_2; datalines; AGREE, NEUTRAL DISAGREE SOMEWHAT, DISAGREE AGREE SOMEWHAT, AGREE ; run; data want; set have; array q q10_1 - q10_2 ; array rq rq10_1 - rq10_2 ; do i = 1 to dim(q); rq[i] = input(q[i],likert.); end; drop i; run;
You could put as many variables on the Q array as there are coded alike just make sure there is a corresponding value in the R array.
Lining things up as I did in the editor may help with that though I would would start with copy the like with the array q and then use the editor search and replace to find q and replace with rq.