close× Call Us
close×
Title The exit-voice trade-off and the decline of public services
Post date 07/09/2018
C1 Background and Explanation of Rationale

Hirschman’s exit, voice, and loyalty model, as set out in his eponymous book, poses a negative trade off between voice and exit. When citizens or consumers have more exit options they will less inclined to voice, taking advantage of exit to realize their preferences. Most tests are observational, suggesting the current experiment. This uses a survey-based approach, presenting respondents with scenarios to examine their voice and exit options. The example respondents are thinking about is health services in the UK, the location for the study. Respondents are given the scenarios of no choice of local health service practitioner (GP) (T1), some choice of GP (T2), and five choices of GP (T3).

C2 What are the hypotheses to be tested?

Randomisation: Groups T1 = no choice, T2= one of two, T3 = one of five GPS

H1: When experiencing service failure, respondents will voice to the extent they experience lack of choice (simplified Hirschman) T1>T2>T3

H2: When experiencing service failure, respondents with no choice will voice less than those who have limited choice, while respondents who experience a wider range of choices will voice less than those who have limited choice (full Hirschman) T1<T2>T3

H3: The reduction in voice in response to exit options is greater for collective voice (with other patients) than individual voice in all conditions (except for T1 comparison with T2 in H2)

H4A: Prior length of time with an actual GP (loyalty) reduces the effect of the choice option on voice (as a moderator in H1 and H2).

H4B: Lower switching behaviour in other contexts (general loyalty) reduces the effect of the choice option on voice (as a moderator in H1 and H2)

H5: The perceived effectiveness of voice will increase the greater the exit options that are available T1<T2<T3

There is a second randomisation toward the end of the survey, which is R1, no response from GP, with R2 as a response from GP, which tests responsiveness after voice or not voicing following a decline in service quality.

H6: If there is a response to the service failure by the GP then satisfaction at the end of the scenario is higher than if there is no response

H7: As a moderator of H6, the difference in satisfaction will be greater for those participants that voiced (in any form) compared to those that did not voice.

H8: If there is a response to the service failure by the GP then exit is lower than if there is no response R1<R2 (for those in T2 and T3 only)

H9: As a moderator of H8, the difference in exit will be greater for those participants that voiced compared to those that did not voice (for those in T2 and T3 only)

C3 How will these hypotheses be tested? *

Two-sided t-tests

OLS regression with and without covariates (see PAP)

C4 Country UK
C5 Scale (# of Units) 3000 people
C6 Was a power analysis conducted prior to data collection? Yes
C7 Has this research received Insitutional Review Board (IRB) or ethics committee approval? Yes
C8 IRB Number 201718-179
C9 Date of IRB Approval 6th July 2018
C10 Will the intervention be implemented by the researcher or a third party? Survation Ltd
C11 Did any of the research team receive remuneration from the implementing agency for taking part in this research? No
C12 If relevant, is there an advance agreement with the implementation group that all results can be published? Yes
C13 JEL Classification(s) not provided by authors