FW: Questionnaire

Mitchelle,

Since my statistics is rusted , you may want to get Aneree’s suggestions examined by someone who is good at it

Pl forward your latest tabulated LIST ( along with headings for 5 columns for scoring ) to Akshay

Hcp

Cc: Aneree : thanks

From: Aneree Parekh [mailto:aneree92@gmail.com]
Sent: 20 July 2017 16:06
To: Hemen Parekh
Subject: Questionnaire

Hi Dadu,

Here are some quick reads on internal and construct validity of a questionnaire as well as Likert scales which you could implement. For internal validity of the scale a simple Cronbach's alpha can be computed and needs to be done only once, at the beginning. 

As for the statistics to be used while looking at organisation and candidate matches, it would really depend on how you will be collecting or coding the data i.e. will it be a single sum number which you will be collecting or will it be ordinal data. I dont think, for the purpose of your client, you will need to run statistical significance analyses, a diverging bar chart would probably be the best way to represent your data. More importantly, though, I dont think you can run sig. analyses on such a small data set (what I am given to understand about 3P, you at most interview 5 candidates for a position). If you so wish to, we might have to look further into the non-parametric tests available. 

Additionally, this is a good (if not long) read on methodologies of developing a questionnaire.

--

Aneree Parekh

+919867283051