I am a doctoral student from GITAM (Deemed To Be University) pursuing my PhD in the area of Positive Psychology, more specifically in the area of psychological capital in educators in the higher education system. I have a query related to my scale which is a standardized questionnaire adopted from different author. Can we use different likert scale. I have used 2 scales with 6 point one with 5 point and another one with 7 point likert scale. Will it make any difference in analysis?
I request you to kindly reply in the affirmative. Please do let me know if there is any additional information I can provide that would be useful. Look forward for a positive response and extended support at the earliest.
Thanking You,
Vasudha Singh
Doctoral Scholar (Positive Psychology)
Hi Vasudha,
The different number of response options in the measures in your questionnaire is probably a good thing. When people have to response only to one type of scale (e.g., 5 points on every item) their attention is more likely to wander than if some scales are 5 point, others 6 or 7.
The number of items in a scale and points on the response scale will lead to different possible minimum and maximum scores on the scales if you sum up people;'s scores. If you want to know whether the mean score on one scale is higher than another then you need to standardise the scores somehow. However, if you are looking at the relationships between the variables measured on the questionnaires statistics like Pearson's correlation standardize the scores within the calculation of the correlation coefficient. in short, if you are looking at correlations and relationships then the number of response options between 5 and 7 will not matter.
I attach our paper on psychological capital so you see which scales we used:
I agree with Professor Curtis, but to make the rule general
In general, we do not choose the scale in a random way. This always depends on many factors, among them the finale use, especially the analysis.
Always avoid changing the scales in the same questionnaire if it is not necessary, and large scales are always better (of course according to the objectives of the study)
Good luck
Please remember that Likert scales are NOT continuous, but are ordinal. "Gaps" between possible responses for a 5 point scale may be different than gaps for a 7 point scale. Additionally, a 6-point scale has no middle point.
Also, again remember that on a Likert scale values between the "whole number choices" are not really possible. That is, an individual can not give something a score of say 3.25 or 3.5 when the choices when the Likert scale contains 1,2,3,4,5. Therefore, interpreting a mean or average between the whole numbers can be misleading. Medians or modes might be the better bets.
- Badges
- Science topic
- Similar topics
- Psychology