I am just wondering, should all the questions which are used in the qualitative observational survey (i.e. KAP studies) needs to be validated and check for their reliability all the time?
I know, if a survey includes the usage of a specific questionnaire, then it has to be validated and checked for its reliability ( like Cronbach's alpha). But what about cross-sectional surveys which just deals with very few questions related to some specific topic like only beliefs or opinions etc.
Dear Sai,
Firstly, this kind of validation/reliability analysis only relates to quantitative data, not qualitative data. Secondly, there is nothing wrong with using data from single questions in a quantitative analysis. Their data type will be nominal, ordinal or scale. However, it is still important to try to ensure that your data is valid, i.e. that it measures what it is supposed to be measuring. This could be achieved through a pilot study which includes interviews or feedback on the instrument.
Reliability analysis is a general name for dimension reduction techniques, such as Cronbach's alpha. It is only relevant when you are trying to reduce the number of variables in your analysis: you must combine at least three together to create a "scale" - see my study guides for more information.
- Badges
- Science topic
- Similar topics
- Social Psychology
- Attitude