Discussion: View Thread

Dear colleague,

We surveyed companies (HR managers) and found low reporting of company level turnover. Here is the footnote we added to the manuscript that  we are still working on:

Low reporting of turnover rates is not unusual. We reviewed 28 studies considered by the meta-analysis conducted by Park and Shaw (2013). These studies all used key informants to obtain employee turnover rates in multi-industry, multi-organization samples. In most, survey and item non-responses were not distinguished, making it impossible to determine the percentage of organizations not providing turnover rates. A notable exception was found in Guthrie (2001) who reported fewer responses on "more sensitive data" (i.e., sales and turnover) that reduced sample size from 190 to 164 organizations (p. 182). Messersmith and Guthrie (2010) received 215 responses from which were removed 25 that did not match their initial selection criteria. The regression analysis they conducted with turnover as a dependent variable had a sample of 108 firms (57% of the analytical sample). Trevor and Nyberg (2008) noted that their multivariate turnover analyses were conducted with only 60% of the initial sample due to missing data. Finally, the sample size reported by Yanadori and Kato (2007) dropped by close to half because of missing values on company turnover and employee tenure variables. Taken together, these studies suggest item non-responses specific to turnover significantly reduced sample sizes by 14% (Guthrie, 2001), 43% (Messersmith & Guthrie, 2010), 40% (Trevor & Nyberg; 2008), and 50% (Yanadori & Kato, 2007).

Let me know if this turns out to be a challenge in your student's research.

All the best,

Victor

------------------------------
Victor Haines
University of Montreal
Montreal QC
(514) 343-7317
------------------------------