We also recorded the identity of the respondent(s), noting concerns about the questionable reliability
ratings from using single-source respondents, whose perceptions are particularly prone to „noise‟ and bias
(Gerhart et al., 2000b; Ichniowski et al., 1996; Purcell, 1999). 63 articles used single raters, of which 26
used a single HR respondent, a matter of some consternation to Osterman (1994: 174) who expressed his
scepticism about the level of awareness and impartiality of the typical HR manager as to her/his
organisation‟s people management processes (also found by West et al., 2002: 1307). 41 articles used
multi-raters (i.e. several respondents per unit of analysis from within a certain sub-population, such as
different managers – e.g. Batt, 1999). 32 articles used multi-actors (i.e. respondents from different sub-
populations per unit of analysis, such as a manager and an employee representative - e.g. Guest and
Peccei, 2001