I'm conducting a masters' research online survey study and am using a range of scales measuring continuous variables e.g. IRI empathic concern and perspective taking (Davis, 1983). I will eventually analyse correlations between 6 variables, and also a multiple regression.
I've found research suggesting item order should be randomised to help eliminate order effects.
However, so many study articles I've read don't mention whether they have randomised or not, so it's hard to say if this is best practice.
Are there any reasons not to randomise item order?
@Asha Bhatt
I would suggest you not to randomise your items in order to improve the reliability of the responses. Rather, brief a little about your variable before its questions/statements in order to create psychology separation in the mind of the respondents so that they can switch off from the previous variable while responding for the next one. For further clarification read the attached manuscript on common method bias.
Dear Asha Bhatt ,
I think there may be two competing issues here: randomisation in order to improve reliability and keeping scales coherent in order to ease completing.
Randomizing the order of scale items is quite rare in moderation survey research. Unless you have some strong reason for doing so, I would avoid it.
Since you are adopting the scales from another researcher, you should keep each scale structure as it was. However, you could simplify the wordings of some items to suit your context, if necessary. Doing so entails expert validation and instrument piloting. Remember, you might need permission to use and reproduce the original instrument. Here is a good read.
Alfred R. Neumann Library (UHCL). (2021, March 16). Research guides: Dissertation format and submission: Getting survey permissions. Research Guides at University of Houston-Clear Lake. https://uhcl.libguides.com/dissertation/permissions
- Badges
- Science topic
- Similar topics
- Psychology