I believe collecting data from multiple respondents from a single organization when the unit of analysis is organization, is not possible. You may take certain positions in all organizations, as the Human Resource Manager, the Finance manager, the Procurement Manager....

I totally disagree with Mahfuz Judeh. The more people you contact in an organization you study, the better. The real understanding of organizational issues appears when all levels are contacted -- from the CEO to shop-floor managers and workers. You can combine interviews of executives with surveys with shop-floor managers and workers.

Enclosed please find an example of the study based on a dozen interviews in the same organization, both in its headquarters and in the major subsidiary.

Well you can, in general, aggregate answers of multiple respondents to the organizational level. For instance, if you want to use the organizational climate as variable, you might ask multiple respondents per company and aggregate their answers. However, you need to run some tests on variance within and across firms first, in order to determine whether your specific setting allows for aggregation (calculate ICC to do so). I'd suggest you to read some literature on multilevel modeling (e.g., Muthen & Satorra, 1995) to familiarize yourself with the required tests and procedures.

Dear Waqar,

Enclosed please fine a few old examples how multiple responses led to the analysis at organizational level. There was the same sample (20 companies) approached annually in 1994-1996.

Here is another example of a paper based on responses of a large number of respondents from the limited set of organizations.

As someone external to the organizations that I study, I rely on a partnership with an internal expert(s) to help me select study participants to get a good cross section of the organization or groups under study. One of the things I request are participants who can provide constructive input, a balance of the good and problem areas. I always find it interesting that the internal experts always know exactly who these people are.

An excellent note by Barry Rubin! The only problem is to find that "internal expert". Usually they can be found in PR departments.

:) Yes, Igor, you have a great point! The challenge is to get an internal expert-partner mature enough to understand that the only way to help an organization get healthier is by dealing with a balance of the strengths and problem areas. They have to see bringing the problem areas to light as something constructive and healthy. I have had partners like these -- some from HR, some from Production -- in my most successful projects. However, they seem to be ambitious and move up to other roles! Look, we have to work with the best conditions we can get -- enough to get results. Human perfection isn't an option.

@ Waqar Rasool : are you designing a qualitative study in which you will be using a small sample of managers or department heads, for example? This would seem appropriate for a single organization as small samples are recommended for qualitative studies. However, for quantitative studies, large samples that are scientifically calculated based on the type of statistical method you identified are recommended to avoid Type 1 and Type 2 errors. You would not have a robust quantitative study with a small sample from a small organization. Just my thoughts...

Thank you so much all of you Mahfuz Judeh Igor Gurkov Nadine Kammerlander Barry Rubin for giving such kind of productive feedback. Actually i am considering quantitative approach Ruth Maria Tappin for my research work. I am working on sequential mediation model and for the data collection i choose knowledge intensive industries including finance, insurrance, education, health and telecommunication in Pakistan. By considerinng key informant technique for data collection. Now i have to decide whether i will choose only one respondent or is it possible that i can consider multiple respondents to fill the survey questionnaire. Because it is very difficult to collect data from the CEOs. Now please guide me in this regard. Thanks

Dear Mr Waqar Rasool I agree with Nadine Kammerlander and suggest you to read the reference given below. Though it is not a common practice of researchers when unit of analysis is Organization, it is recommended in recent literature to use more than one informants or respondents (in quantitative survey study context in this case) to have better reflection of study constructs.

Please consult this reference given below:

(Bou‐Llusar, Beltrán‐Martín, Roca‐Puig, & Escrig‐Tena, 2016); Single‐and Multiple‐Informant Research Designs to Examine the Human Resource Management− Performance Relationship. British Journal of Management.

Good reference!

I.G.

Tanks Prof. Igor Gurkov for your endorsement.

Thank you very much, it was realy an honor for me as a research student to having such productive feedback from all of you. It helps me alot. Thanks again everyone of you.