Hi guys :)
I would really appreciate your help.
I am busy writing my masters thesis. For my survey, I have a 3x2 design. For each condition, I have two questions (that are the same).
When exporting my data from Qualtrics and uploading it to SPSS I got quite confused (it's been a while since I last used SPSS haha so that's also probably why). Attached is a screenshot of my data view in SPSS (and the two questions per condition), if you look at the top I suddenly have so many conditions? I think maybe it's also for each question in the different conditions?
How do I recode this to make it easier to analyse my data?
Not sure if any of this is making sense but any help would be amazing!!
Thank you!
Hello Britt,
The data screenshot makes it appear as though cases' responses are clustered into six variable sets (which, for a 3 x 2 design, seems plausible). Of course, it would make more sense if the response variables were the same for all cases, and cases had two associated variables to indicate which level of the "3"-level factor they belonged to and which level of the "2"-level factor they belonged to.
If the survey logic was set up to shunt respondents to one of six item sets based on their condition, then the importing would look just like what you are seeing. This isn't ideal, of course, but maybe it was the easiest to implement in qualtrics. (I can't help but believe that the qualtrics logic could be set up to record the data in the preferred way.)
So, what can you do?
1. Set up a set of recode statements to save responses for the first cluster you see on the page (of the six apparent clusters) to a new, common, seventh cluster. Do this in syntax, so that you can repeat below, with only minor edits.
2. Repeat this for all six clusters. (You might need to preface each run with a select if statement, so that you are saving the responses only for cases in that subset/cluster).
3. Use the new, seventh cluster as your response set.
Good luck with your work.
Hi David,
Thank you for your answer!
Unfortunately I have never worked with syntax before in SPSS.
My supervisor advised this:
"You need to make one dataset, and the outcome and the (items of) mediator scales should become one variable each time. So just put them under each other if they are not already. The conditions then become dummy variables. So if 1-30 are for condition one these get a 1 and 0 otherwise, if 31-52 are then condition 2 these get a 1 for condition 2. Later you can also make dummies of the variables underlying the condition. And then you first test the mediator and DV based on the groups. And later do a regression where the dummies are the IV and moderator."
However I am unsure how to follow these steps in SPSS. And there seem to be no tutorials online covering this topic (how to recode randomised conditions from Qualtrics in SPSS)
If you have any more tips to make it actionable based on what my supervisor said above that would be greatly appreciated.
Thanks,
Britt
- Badges
- Science topic
- Similar topics
- Statistics
- Survey