You're looking at Less Wrong's discussion board. This includes all posts, including those that haven't been promoted to the front page yet. For more information, see About Less Wrong.

thomblake comments on 2012 Less Wrong Census Survey: Call For Critiques/Questions - Less Wrong Discussion

20 Post author: Yvain 19 October 2012 01:12AM

You are viewing a comment permalink. View the original post to see all comments and the full post content.

Article Navigation

Comments (479)
Tags: meta

Comments (479)

You are viewing a single comment's thread. Show more comments above.

Comment author: thomblake 19 October 2012 03:12:50PM 2 points [-]

I will probably not be implementing any suggestion that requests a multi-checkbox style question...

Wow, the way Google Docs presents that data is really annoying for analysis. Normally, each option would be given its own column in the data (as though it were a 'Yes/No' like you suggested).

A bit of code from here for splitting a csv field in SPSS:

DEFINE !parse (var=!TOKENS(1) /nbval=!TOKENS(1))
COMPUTE !var=CONCAT(RTRIM(!var),';').
STRING #str(A8).
VECTOR !var (!nbval F8.0).
COMPUTE #beg=1.
LOOP #cnt=1 TO !nbval.
+COMPUTE #str=SUBSTR(!var,#beg).
+COMPUTE #end=INDEX(#str,';')-1.
+DO IF #end=-1.
+ BREAK.
+END IF.
+COMPUTE !var(#cnt)=NUMBER(SUBSTR(#str,1,#end),F8.0).
+COMPUTE #beg=#beg+#end+1.
END LOOP IF #end=-1.
EXECUTE.
!ENDDEFINE.
* Call the macro.
!parse var=c254 nbval=5.
!parse var=c256 nbval=5.

Replace the semicolons above with commas, and it should work. Note that nbval is the number of possible options, and this code will give warnings if fewer are selected.

Comment author: thomblake 19 October 2012 03:17:42PM 4 points [-]

Actually, what I'd recommend doing for real, is moving the data to a better spreadsheet program like Open Office, doing the split programmatically there (as suggested elsewhere), and then importing the data into SPSS.

Comment author: ChristianKl 19 October 2012 07:05:40PM 0 points [-]

SPSS is a software build for doing statistics. There should be a way in SPSS to do this that's better than switching to a different software.

Comment author: thomblake 19 October 2012 07:22:21PM 5 points [-]

It's software built for doing statistics a long time ago. And it makes certain assumptions about the sorts of data you're going to be giving it. This is a solution for dealing with Google's silly nonstandard way of representing multi-punch lists, not a solution to a statistical problem. It's totally reasonable to process the raw data using some other program or script before sending it to SPSS - it's what the pros do anyway.

Comment author: Kawoomba 19 October 2012 07:18:16PM 3 points [-]

R is a free - and imo better - alternative.

Comment author: ChristianKl 19 October 2012 07:26:45PM 1 point [-]

I'm personally using R myself. That's why I can't imagine it to be hard is SPSS because. If Yvain has however spent time learning SPSS it might be costly to relearn things in R.

Powered by Reddit Powered by Reddit