Dear Jyoti Ghosal ,

I think it depends on what type of data you are collecting and whether you are planning to combine several questions together to represent the same construct. If you have collected ordinal or numerical responses and wish to combine several answers together to represent the same construct then the best technique to establish and validate your scales is Exploratory Factor Analysis (although it is actually a whole spectrum of different techniques). Please see my study guide for more information.

Peter Samuels may have covered this in his study guide BUT in my humble opinion Science isn't about 'what can I do to show that it is right' but 'how can I get it right'. You will actually have to go through the test-analyse-retest cycle, using a wide variety of psychometric techniques until you arrive at a measuring tool which is robust enough to stand the scrutiny of your peers. I do not doubt for a moment that Peter would disagree that creating a research tool is hard work and there are no short easy cuts.

In addition to Peter's study guide, you may wish to consult my question-and-answer web site at uxp.ie/query .

From what I read you have started well and you are asking the right questions!

@Jyoti below are few studies on scale development and validation and I hope they would really help your cause. Wishing you best for you research.

https://www.emerald.com/insight/content/doi/10.1108/RAUSP-05-2019-0098/full/html

https://www.frontiersin.org/articles/10.3389/fpsyg.2020.00799/full#:~:text=After%20the%20scale%20has%20been,predictive%2C%20concurrent%2C%20and%20discriminant.

https://methods.sagepub.com/book/public-health-research-methods/i1816.xml

Hi Jyoti Ghosal,

The surveys are validated by providing pieces of evidence of face and content validity before the survey is administered. After the administration of the survey, internal consistency reliability can be used, and make sure you have 5-10 responses per question in the survey (this is a well know practice followed).