Hi all,
Just had a quick question about alternating survey response options and wanted to get some insight into the practice of re-coding said response options.
I'm currently working with survey data whose response options tend to alternate between:
Always
Often
Sometimes
Rarely
Never
and
Strongly agree
Agree
Neutral
Disagree
Strongly disagree
I was just wondering whether (for analyses purposes) I could treat the two response scales synonymously, i.e, recode Always and Strongly Agree as 1's, Often and Agree as 2's, Sometimes and Neutral as 3, etc...
Dear Darko,
Though this doesn't sound irrational but the problem is, for a responder the meaning of sometimes and neutral or often and agree may not be same. So, conservatively its not suggested.
Also, if it is in your thesis/ part of it, the board will get100 options to question your data collection.
Could you clarify what kind of questions the scales was used for?
If they refer to behaviour (how often do you do X? always, ..., never) then it's not very comparable with feelings/opinions questions. Someone might rarely do X but still feel very positively about it and visa versa. However if the always to never scale questions also refer to feelings/opinions just like the agree to disagree questions you have a stronger case.
In my opinion - this would not be valid.
Participants interpreted items in a certain way when they answered them ... if you simply re-code and change their responses, your data will be invalid.
Stick to the original data - run your descriptive analysis - you will certainly find patterns between the two different types of questions - this will certainly be of interest to the reader of your research.
Hi all, thank you for all of the input! It's greatly appreciated.
Jolien van Dijk , a sample question that utilizes the two separate response options would be:
Likes to question things they're learning (Always, Often, Sometimes, Rarely, Never)
and
The skills being learned will help me later on life (Strongly Agree, Agree, Neutral, Disagree, Strongly Agree).
I've just been coding these as 1-5, rather than focusing on value labels themselves.
- Badges
- Science topic
- Similar topics
- Statistics
- Survey