Calculating standard deviation

[Sowme Saha]

Hi all,
I was hoping to get some guidance on a problem that I'm facing.

The data that I am working with looks something like this:

Responses	Questions
	1.1(A)	1.1(B)	1.1(C)	1.2(A)	1.2(B)	1.2(C)	2.1(A)	2.1(B)
Sally	4	5	4	3	18	6	3	2
Joe	4	6	7	7	7	3	35	3
Allen	2	7	3	8	5	2	4	8

I have far more responses(>1000) than those listed here.
I need to find the standard deviation of all entries to questions ending in (A) for each respondent. So for Sally, I need to find the std dev for the values 4,3,3
Any ideas on how I can achieve that?

Thanks!

[Bruce Weaver]

Hello Sowme. The column headings in your table above are not legal variable names. What are the actual variable names? If the actual variable names end in A, try something like this:

egen sdAvars = rowsd(*A)

For more info:

help egen

Then Ctrl-F and search for rowsd.

HTH.

p.s. - Remember that Stata is case-sensitive, so a and A are treated differently.

[Sowme Saha]

Thanks Bruce that helps. One followup question -
suppose I wanted to include all variables ending in A but not 1.2(A) in the std dev calculation, how would I do that?

[Bruce Weaver]

Thanks Bruce that helps. One followup question -
suppose I wanted to include all variables ending in A but not 1.2(A) in the std dev calculation, how would I do that?

I don't know a wildcard shortcut for that one. Maybe one of the veteran members of this forum knows one. If not, you'll have to list all of the variables inside the brackets.

[Carole J. Wilson]

1.2(A) is not a legal variable name in Stata. If you have some number of variables ending in "A" and you want to get the row standard deviation for all of those variable except for a variable with the name var_1_2_a.

unab avars: *a
local takeout var_1_2_a
local vlist: list avars - takeout
egen avar_sd=rowsd(`vlist')

[Sowme Saha]

Responses	Questions
	time1period1eat	time1period1drink	time1period1play	time1period2eat	time1period2drink	time1period2play	time2period1eat	time2period1drink	time2period1play
Sally	4	5	4	3	18	6	3	2	2
Joe	4	6	7	7	7	3	35	3	6
Allen	2	7	3	8	5	2	4	8	1

Please see above variable names.

Essentially , for each respondent, I need to find the std dev of all variables that contain 'eat' except for the variable 'time2period1eat'.
Please note that there are more variables than the ones listed here.

Thanks!

[Carole J. Wilson]

In the above code:
Replace *a with *eat in the first line.
Replace var_1_2_a with time2period1eat.

[Bruce Weaver]

Re #5, I was unaware of the -unab- command. Thanks for the pointer, Carole. For others who may not have known about it:

help unab

[Sowme Saha]

Thanks Bruce and Carole, it works!