Psych Safety, by Iterum

Predicting Exit, Voice, Loyalty, and Neglect

Withey & Cooper (1989) · Loyalty that looks like entrapment

Withey & Cooper · Primary Research · Voice & Silence · Administrative Science Quarterly · 1989 · Paywalled

Open this neighbourhood on the live map →

Read with care: Some findings did not replicate
Internal consistency was low for voice (.41) and neglect (.55), and some findings failed to replicate in the paper's second, smaller sample.

Michael Withey and William Cooper ask when dissatisfied employees respond with exit, voice, loyalty or neglect, the typology Rusbult and colleagues built from Hirschman (1970) by adding neglect. In their theory of action people weigh the cost of an action, its efficacy and the attractiveness of the setting. Costs were exit costs (skill specificity, sunk costs, investment) and voice costs, including the likelihood of punitive responses to attempts at change; efficacy was prior satisfaction, belief that improvement is possible and locus of control; attractiveness was organisational commitment and better alternatives.

Two longitudinal samples were surveyed in 1984, with responses measured six months after the predictors. Of 915 Queen's University commerce graduates, 460 replied (50.3 per cent) and 266 who stayed were analysed after exclusions, the 37 who had left being set aside. In a mid-size Ontario accounting firm, 124 employees answered twice and 40 were interviewed.

Exit was the most consistently predicted response, rising with low sunk costs, high voice costs, dissatisfaction, little hope of improvement, low commitment and a better alternative (adjusted R² .26 among graduates; 23 per cent of variance in the firm). For voice the model got the signs right but 'didn't explain much voicing' (adjusted R² .05), and in the firm it explained none. Among graduates, loyalty rose with low prior satisfaction, pessimism about improvement, an external locus of control and low commitment, three of these opposite to prediction, and the same variables predicted neglect, which differed only in responding to a better alternative. Interviews portrayed the loyalist 'as an entrapped person'. As the loyalty items asked about waiting patiently and saying nothing, they conclude 'we got the loyalty results we deserved'.

They name self-report as a limit that the longitudinal design reduces but does not remove. Internal consistency was low for voice (.41) and neglect (.55), and some findings failed to replicate in the smaller firm, which they say suggests some graduate findings 'are unreliable'.

Several unhappy accounting staff did not voice because they did not think any good would come of it, and saw supervisors as too hard to find and too busy to listen, barriers 'not captured by the voice-cost construct'. The paper does not use the term psychological safety. The bridge is ours: we read those interviews as futility and absent listeners keeping people quiet alongside the voice costs the survey measured, which went with less voicing among graduates only, and the paper's loyalty, measured as waiting and saying nothing, as silence that resembled entrapment more than allegiance.

Withey, M.J. and Cooper, W.H. (1989) 'Predicting exit, voice, loyalty, and neglect', Administrative Science Quarterly, 34(4), pp. 521–539.

Save this paper to a bibliography, keep notes on it, and export citations. Create a free account →

Related psychsafety.com articles (5)

Related papers (15)

View the source paper →