Table 2 Main study limitations and risk of biases in the 83 studies reviewed

Manlove [111]

X

X

1993

Deary et al. [48]

X

The use of zero-order statistics for the evaluation of multifactorial constructs, such as psychological issues, has been recognized as being inappropriate

1996

Deary et al. [112]

1996

Mills et al. [60]

X

X

X

The exclusion of the personality measure at Time 2 did not allow for a complete analysis of the effects of personality differences on burnout dimensions

1998

Zellars et al. [113]

X

X

2000

Zellars et al. [53]

X

X

X

2001

De Vries et al. [114]

2002

McManus et al. [50]

2004

Zellars et al. [52]

X

X

X

X

2004

Cano-García et al. [115]

2005

Burke et al. [116]

X

X

X

2006

Goddard et al. [57]

X

X

Participants may have different characteristics from those of those who did not participate

No comparison with a control group

2006

Langelaan et al. [117]

X

X

2006

Mostert et al. [118]

X

X

2006

Bahner et al. [119]

X

X

The focus of the study was on the male batterer intervention program in the context of heterosexual relationships, not that of same-sex relationships

It is not known whether the results may reflect the different nature and structure of the working environments of the batterer intervention programs

2007

Ghorpade et al. [120]

2007

Kim et al. [121]

2007

Teven [122]

X

X

X

The data comes from a sample with only the level of education university

Respondents may have tried fatigue because many of the measures overlap conceptually

2007

Leon et al. [123]

X

X

X

There are missing data and differences between the samples for age and education

No specific job characteristics to RTCs were explored that might be associated with the client presentations measured

It is possible that this study's moderation findings will not generalize to other samples of front line staff

2008

Chung et al. [124]

Challenging behavior has been measured in terms of the level of challenging behavior of the clients with whom they worked most intensively, and this does not accurately or reliably reflect exposure to challenging behavior

No data on staff views of their job, on assaults or witnessed violence by staff, or on the kind of recent significant life events of the staff

2009

De Hoogh et al. [125]

X

X

In Sample 1 ratings of independent and dependent variables were provided by the same subordinates

In Sample 2 the selection of raters may have selected employees who share attitudes and opinions

2009

Gandoy-Crego et al. [126]

X

2009

Kim et al. [127]

X

X

X

Focus limited to individual differences

2009

Taormina et al. [128]

X

X

X

2009

Barford et al. [129]

X

X

X

X

Different methods of data collection were used

2010

Perry et al. [51]

X

X

X

The distinct contexts across the two samples and lower variance across variables in Sample 1 may have affected the results

2010

Ghorpade et al. [130]

2011

Hudek-Knežević et al. [59]

X

X

Burnout at Time 1 was not measured

Sample attrition due to drop-out at Time 2

Significant differences (age, agreeableness, and continuance commitment) were found between the sample included and the one that in both Times and the one that dropped out

2011

Salami, [131]

X

X

2011

Sterud et al. [61]

2011

Armon et al. [54]

The findings could be biased because of the "healthy worker effect"

Longitudinal design was based on only two waves of measurement

The personality scale only return a total score and do not evaluate the influence of the personality variables on each of the burnout factors

The possibility that the factors of the personality may interact in affecting burnout levels has not been tested

2012

Zimmerman et al. [132]

X

The sample was composed exclusively of high-level professionals

Only neuroticism and extraversion and not all personality traits were evaluated

2012

De la Fuente Solana et al. [133]

X

2013

Garbarino et al. [134]

X

X

X

2013

Hurt et al. [135]

2013

Lin et al. [136]

X

X

The sample was relatively young

Questionnaires used are not popular therefore the comparison with other studies (also of other nationalities) may be limited

The employees who did not respond to our questionnaire might have been those who were busier

2013

Gan et al. [56]

X

X

X

Could not examine the reciprocal relationship (burnout/engagement to demands/resources) using panel data

2014

Reinke et al. [137]

X

X

X

2014

Taycan et al. [138]

X

Lack of an urban physician sample for comparison

2014

Yilmaz, [139]

2014

Cañadas-De la Fuente et al. [140]

X

X

X

2015

Ang et al. [142]

X

X

X

2015

Iorga et al. [143]

X

X

2016

Vaulerin et al. [144]

X

2016

Zhou et al. [145]

X

X

Some scales have been adapted to the sample under study and their construct validity and temporal stability have not been confirmed

The questionnaire was completed on workdays and may have caused additional stress for the firefighters

2016

De la Fuente-Solana et al. [146]

X

X

X

2016

Geuens et al. [147]

X

X

X

2017

Iorga et al. [148]

X

X

X

X

The study was conducted from November 2012 to July 2013 and could lead to different results than a recent one

The focus was only on individual factors, leaving out job-related and organizational factors

2017

Lovell et al. [149]

X

X

X

The access to medical facilities and the small distance from patients who live in urban areas may have diminished the physicians' professional exhaustion

2017

Ntantana et al. [150]

X

X

Methods based on measuring physiological parameters are not objective

2017

Al Shbail et al. [151]

X

X

X

Causal factors have not been investigated

2017

Bergmüller et al. [152]

X

X

Only companies that were part of the Amman Stock Exchange (ASE) were included in the study and not the others

All types of shareholding companies were included in the study, regardless of sub-sectors

2018

Bianchi et al. [153]

2018

Bianchi, [19]

X

X

X

It was used only one measure of burnout

The role of organizational and work-contextualized factors may have been overemphasized in burnout research, to the detriment of the role of generic dispositional factors

Burnout's discriminant validity is unsatisfactory

2018

Harizanova et al. [49]

X

X

An exhaustion-centered measure of burnout was used

Only neuroticism and extraversion and not all personality traits were evaluated

2018

Hildenbrand et al. [58]

X

2018

Iorga et al. [154]

X

X

X

The correlation between thriving and burnout at Time 1 and Time 2 was rather high

Despite large groups of employees sharing the same supervisor, it was not possible to collect information on group composition and therefore take into account the variance due to team membership

2018

Tang et al. [155]

X

X

2018

Tatalović Vorkapić et al. [156]

X

X

The psychological distress of the participants was not measured and also the fitting degrees of both logistic regression modes are low, suggesting that there are other unknown risk factors

2018

Yao et al. [157]

X

X

2018

Zaninotto et al. [158]

The participants of the sample, coming from the municipal hospitals, had heavy patient loads and therefore more exposed to burnout

2018

Bahadori et al. [159]

X

X

A large number of correlations increases the risk of type 1 errors

2018

Brown et al. [160]

X

X

X

2019

Castillo-Gualda et al. [62]

X

X

X

2019

De la Fuente-Solana et al. [161]

X¹

X

X²

X²

X

2019

De Looff et al. [55]

X

X

X

2019

Farfán et al. [162]

X

X

Methods based on measuring physiological parameters are not objective

2019

Khedhaouria et al. [163]

X

X

X

X

2019

Pérez-Fuentes et al. [17]

X

X

Organizational factors that reduce stress, the effect of technostress, and personality traits on job burnout among employees, nor cases that could have specific antecedent conditions predisposing job burnout were not evaluated

2019

Ye et al. [164]

Some information is lost due to some scales, which are too short to adequately evaluate the personality construct, or they only return a total burnout score and do not evaluate the influence of the personality variables on each of the burnout factors

2019

Banasiewicz et al. [165]

X

X

The Baron and Kenny method to examine the mechanism of the mediator is too simple

2019

Bhowmick et al. [166]

X

The questionnaire contained closed-ended type questions

The questionnaire required a long duration of completion

Pregnancy terminations are performed only at several facilities in Poland

2020

De Vine et al. [167]

X

X

X

X

X

X

X

2020

Dionigi, [168]

X

X

2020

Farfán et al. [169]

X

X

X

The predictors evaluated accounted for only a part of the burnout dimensions

2020

Liu et al. [170]

X

X

X

Sample with high concentration of people with a high education level

2020

Mahoney et al. [171]

X

X

The study focuses on negative workplace gossip and does not include positive workplace gossip

The scale of negative workplace gossip is based on frequency, but the perception of negative workplace gossip may have different influences

The mean score for negative workplace gossip is relatively lower than those of other studies with Chinese samples

2020

Malka et al. [172]

2020

Tasic et al. [173]

X

The issue of work supervisors who wanted to leave the job was not evaluated

Being a quantitative study, it lacks the interpretive perspective of the participants

It is not clear to what degree participants related to their role as fieldwork supervisor compared to their experiences workers in general

2020

Bianchi et al. [174]

2020

Fuente-Solana et al. [175]

X

X

X

X

X

X

X

French sample was considerably larger than Spanish and Swiss samples

2021

Fuente–Solana et al. [176]

X

X

X

Methods based on measuring physiological parameters are not objective

There was the impression that a sample with a certain level of burnout was more prone to participate

2021

Manlove [111]

X

X

X

X

Methods based on measuring physiological parameters are not objective

2021