Human resource management training of supervisors for improving health and well‐being of employees
Abstract
This is a protocol for a Cochrane Review (Intervention). The objectives are as follows:
To assess the effect of HRM training for supervisors on employees' psycho‐mental stress, absenteeism, and well‐being.
Background
Description of the condition
The workplace and especially the psychosocial work environment are important determinants of the health and well‐being of employees (Brunner 2006;Joyce 2010;Marmot 2006;Marmot 2012). Trends such as increased work pace, more highly skilled jobs, and the increased use of information and telecommunication technology have been placing increasingly higher demands on the mental functions of employees (European Agency for Safety and Health at Work 2007;Nieuwenhuijsen 2010). Psychosocial risk factors in the working environment are associated with higher levels of work stress, which is reported to be experienced by about 22% of European workers (Houtman 2005). In addition, an online‐survey conducted in 2001 in more than 30.000 workers worldwide revealed that globally 26% (range 17‐35%) of workers felt to be under unreasonable work stress (D'Mello 2011).
While successful coping with work stress may have positive effects on performance and quality of life, unsuccessful coping may in the long run lead to coronary heart disease, musculoskeletal problems, adverse health‐related behaviour (e.g. smoking, substance abuse), impaired mental‐health (e.g. anxiety, burnout, depression) or other stress‐related symptoms (Fransson 2012; Heikkilä 2012a; Kivimäki 2012; Kuper 2002; Lang 2012; Siegrist 2006; Stansfeld 2006). Importantly, feeling stressed should not only be seen as a risk factor or intermediate variable alone, but also as a condition of reduced quality of life in itself.
To our knowledge, there are no generalisable data on rates of sick‐leave that could be solely and specifically attributed to work‐related psychomental stress or consecutive stress‐related symptoms. Likewise, it should be taken into consideration that the total rate of absence from work is not always due to illness alone, that is, justifiable sick leave, but also due to unjustifiable absenteeism, which can be defined as the practice of regularly staying away from work without good reason (Darr 2008; Oxford Dictionaries 2013). Therefore, the over‐all rate of absence from work should be seen as a combined measure influenced by a mixture of medical, psychological, and social factors as well as individual work attitudes like withdrawal, commitment, work‐engagement or job satisfaction (Darr 2008;Johns 2007). A survey conducted by the European Union (EU) in 2007 revealed that 23 million people, representing 8.1% of current or former workers reported to have had a work‐related health problem within the preceding year (European Commission 2010;European Commission 2011). It was estimated that these problems resulted in at least 367 million calendar days of sick leave (European Commission 2010; European Commission 2011). The overall rate of sickness absence ranged from about 3.5% to 8.0% within the EU during the 1990s (Eurofound 1997). Where as in 2011 it was 4.7% in Germany (WIdO 2012), 1.8% in the UK (Office for National Statistics 2012), and 3.0% in the USA (Bureau of Labor Statistics 2012). The economic burden caused by absence from work (due to all causes) in the EU countries was estimated between 1.5% and 4% of the gross domestic product (GDP) in the 1990s (Eurofound 1997; Livanos 2010).
After musculoskeletal problems, mental (stress‐related) disorders like burnout, anxiety and depression (14%) have been the second most frequent work‐related health problems in this survey (European Commission 2010;European Commission 2011). In Germany, all mental diseases accounted for about 10% of all sick‐leave days but may also include non stress‐related cases (e.g. personality disorders, substance abuse) (WIdO 2012). This number has increased by nearly 60% since the year 2000 in Germany (WIdO 2012). In particular, it should be mentioned that the rate of sick leave due to burnout syndrome has increased more than 10‐fold in Germany since 2004 (WIdO 2012). In 2011, burnout syndrome caused more than 94 days of sick leave in 1000 people (WIdO 2012).
Description of the intervention
This review will include studies on all human resource management (HRM) training programs that aim to enhance the knowledge, the attitude, the skills and the behaviour of supervisors. HRM training programs are widely used to improve leadership in organisations (Yukl 2012). Training can take many forms, from self‐helping activities to developmental activities (e.g. 360° feedback, coaching), from short workshops to programs that last for a year or more, from programs focused on skills needed in the current position to training preparing managers for promotion to higher positions, or from programs tailored to a company's needs to workshops imparting generic skills (Yukl 2012). We classify the interventions relevant for this review based on the following two dimensions (Table 1).
| Training off‐the‐joba — Self‐help activities (textbook‐based, video‐based, web‐based) — Formal training (face‐to‐face lecture, classroom lecture), case analysis and discussion— Simulations— Role playing — Behavioural role modelling — Developmental assessment centres | Training on‐the‐joba, (real time/things/situations) — Executive Coaching — Mentoring — (360°) Feedback — Job rotations — Action learning | |
| Supervisor‐Employee Interaction — Communication (e.g. active listening, clear instructions, role clarity, clear and ethical organisational and personal goals, etc.) — Justice (interactional) — Recognition and immaterial reward (appreciation, respect, etc.) — Supervisory support (emotional) — Leadership style | (A1) e.g. web‐based teaching of how to improve active listening, role‐playing simulation for improving active listening skills | (A2) e.g. personal coaching concerning active listening while working |
| Design of working environment b — Participation — Justice (procedural and distributive, including material reward) — Autonomy (e.g. working time/schedule) — Decision latitude and control — Supervisory support (instrumental) | (B1) e.g. classroom lecture for improving knowledge on positive health effect of participative leadership, case‐analysis simulation for improving participative leadership behavior | (B2) e.g. 360°‐feedback session for assessing and improving the leaders participative leadership behavior |
(a) For example, combined interventions containing both, an introduction to active listening skills using a video‐based lecture (A1) followed by personal coaching sessions (A2) will be classified as A2. For another example, a short video‐based introduction on active listening (A1) followed by an considerably longer role playing session on participative leadership (B1) will be classified as B1.
(b) Only included if these conditions could (and should) be influenced directly by the targeted supervisor.
-
First, we distinguish between HRM training programs (i) aiming to improve the dyadic interaction (e.g. relationship, communication) between supervisors and employees, and (ii) aiming to improve the capabilities of supervisors to change the psycho‐social working environment in a health‐promoting way.
-
Second, we classify interventions according to their proximity to real life by distinguishing between (i) training off‐the‐job such as in simulated conditions and (ii) training on‐the‐job while working with real people on real problems (adapted from Australian Public Service Commission 2011 and Nohria 2010).
The combination of both dimensions results in four categories of interventions (Table 1). The table presents practical examples of what HRM training programs may look like.
How the intervention might work
The linkage between HRM training for supervisors and health outcomes of employees is rather complex, often indirect and influenced by a plethora of contextual factors (Figure 1). According to Brunner 2006, employee health is related to the social structure and work environment via several psychological factors and health‐related behaviours. Thus, HRM training programs might work by influencing the following two pathways.
Logic model of how HRM training of supervisors may lead to positive changes in employees
-
Reduction of work‐related stressors: Important work‐related psychosocial risk factors and possible sources of work‐related stress (work stressors) are listed in Table 2. Supervisors may influence some of these stressors, but this also depends on the organisational, structural, cultural and other environmental conditions. For example, 360°‐feedback sessions could identify such work‐related stressors as communication deficits between supervisors and employees that manifest as unclear or contradictory instructions. These deficits may be worked through during personal coaching sessions for supervisors.
-
Promotion of work‐related resources: As a dimension of mental health, well‐being has been defined as a state in which every individual realises his or her own potential, can cope with the normal stresses of life, can work productively and fruitfully, and is able to make a contribution to her or his community (WHO 2005). Therefore, it is crucial to encourage positive psychological or salutogenic approaches (Diener 1999; Ryff 1996; WHO 2005), for example by promoting work‐related resources as listed in Table 2. In the absence of any work stressors, work‐related resources themselves are unlikely to significantly influence health status, but when employees are exposed to stressful situations resources may improve the capability of employees to successfully cope with stress. For example, a role‐playing simulation that aims to improve the ability of active listening and giving social support to the employee may lead to improved considerative and supportive behaviour of the supervisor. In turn, the employee's perception of altered supervisor behaviour (being recognised, receiving consideration and support) will act as a resource as described above.
| Work‐related stressors | Work‐related resources |
|
|
|
The list is neither exhaustive nor does it make any claims about completeness. Some factors can be both, a work‐related stressor and a resource depending on whether they are present/pronounced or absent/less pronounced (Cropanzano 2005;Eurofound 2010; Houtman 2007;Kals 2012; Lohmann‐Haislah 2012; Semmer 2010; Stadler 2003;WHO 2013; WHO Regional Office for Europe 2010).
In addition, supervisors may indirectly influence employees' behaviours as a consequence of coping with work stress (Knoll 2011; Siegrist 2006). Behaviours that can be affected include for example smoking, alcohol use, drug abuse, presenteeism (meaning attending work while ill) (Johns 2010), physical activity, diet, and adherence to regular check‐ups or health‐promotion programs (Bamberger 2006; Heikkilä 2012a; Heikkilä 2012b; Kelloway 2010).
Furthermore, supervisors' support of work‐place health programmes is a critical determinant of the success of employee‐focused health promotion interventions in organisations (Kelloway 2010). Concurrently, HRM training programs may also influence several context variables or modifying factors, which in turn influence coping with stress. However, we will not explicitly analyse these complex interactions between HRM training, the behaviour of supervisors and the context in this review. In order to identify the specific effect of HRM training as compared to everything else going on in the complex work environment, we will carefully document intervention components and any relevant contextual information provided.
Why it is important to do this review
As mentioned above, about one quarter of workers who had participated in an EU‐wide and global survey suffered from work‐related stress and are therefore at considerably increased risk for work‐related cardiovascular, musculoskeletal, and mental disorders (Ariens 2001; Belkic 2004; D'Mello 2011; Hoogendoorn 2000; Houtman 2005; Kivimäki 2002; Kivimäki 2012; Lang 2012; Marmot 2006; Stansfeld 2012).
There is consensus that onset of stress, stress consequences and degree of well‐being could be influenced by leadership behaviour (Gregersen 2011; Kuoppala 2008; Nieuwenhuijsen 2010; Nyberg 2005; Skakon 2010). A number of studies have been conducted to analyse leadership behaviours in detail but their predominant purpose was to prove the eligibility of these models to predict work performance or job satisfaction. A part of these studies investigated the impact of leadership training on employee health and well‐being. Recent reviews have been performed in a more or less systematic way but they have included predominantly non‐experimental studies (Gregersen 2011; Kuoppala 2008; Nieuwenhuijsen 2010; Nyberg 2005; Skakon 2010). Two of these reviews contain randomised controlled trials (RCTs) but only one each (Kuoppala 2008; Nyberg 2005). Thus, we can derive only associations rather than true causal relationships between leadership interventions and health outcomes from these reviews. The most recent review of the literature by Tsutsumi 2011 focused on three RCTs and four quasi‐experimental studies. However, the authors did not assess included studies' risk of bias nor did they perform a meta‐analysis of results data.
Objectives
To assess the effect of HRM training for supervisors on employees' psycho‐mental stress, absenteeism, and well‐being.
Methods
Criteria for considering studies for this review
Types of studies
We will include all eligible RCTs and cluster‐RCTs. We will also include controlled before‐after studies (CBAs) but we will not use them for drawing conclusions. We will include CBAs only if they have at least two intervention sites and two control sites and only if the outcome was measured both before and after the intervention (Cochrane EPOC Group 2013b). We will exclude all other study designs. We believe that including CBAs will substantially increase the applicability of our review and its value for potential users.
We will consider studies reported as full‐text, those published as abstract only, and unpublished data.
Types of participants
We will include studies enrolling any type of supervisors (male or female) and their dependently employed subordinates (male or female). For the purpose of this review a supervisor is defined as a person who has the authority to give instructions to at least one subordinate and is held responsible for the work and actions of his or her employees. We will include studies that have been conducted in profit, non‐profit or governmental organisations, that is, in a real working environment. We will exclude studies that have been conducted in special settings like unpaid work, work without contract, freelance work, internships, volunteering or any comparable conditions. We will include studies conducted in mixed settings such as both paid and unpaid work. We will not apply any further restrictions concerning the study population. We will also include all studies conducted only in subgroups or in companies organised other than by line‐management (e.g. by matrix‐management).
Types of interventions
We will include studies comparing HRM training of supervisors (as listed in Table 1) with a passive control group, such as a waiting list or no intervention, or to an active control group receiving an alternative intervention. We will consider the following two categories of HRM training: (i) training off‐the job (e.g. formal face‐to‐face lectures, simulations, role playing), and (ii) supervisor training on‐the‐job (e.g. personal coaching, feedback sessions). We will consider HRM training that aims to change supervisor‐employee‐interaction (e.g. communication skills, providing support, transformational leadership behaviour) or the design of working environment (e.g. justice, participation). We will include interventions regardless of the duration of HRM training (Table 1). We will exclude studies that do not directly target supervisor behaviour but focus on general improvements of work organisation, which are not amenable to direct change by the supervisor (e.g. overall working time or general reward system). Further, we will exclude studies examining interventions that target the health behaviours of employees or supervisors (e.g. stress self‐management programs).
Types of outcome measures
We will include all studies reporting at least one primary outcome of our review as described below. All outcome measurements must have been performed in employees (not in supervisors).
Primary outcomes
We will include studies that have measured the effectiveness of HRM training on:
-
validated measures of psycho‐mental stress such as the Maslach Burnout Inventory or the Perceived Stress Scale (Cohen 1983; Maslach 1996);
-
any estimate of absenteeism;
-
measures of well‐being such as the World Health Organization (WHO) 5‐item Well‐Being Index (WHO‐5), or work‐engagement scales (Psychiatric Research Unit 1998; Schaufeli 2002).
Secondary outcomes
When included studies have measured one or more of the primary outcomes listed above, we will also report results measured as changes in employees' health‐related behaviour such as smoking, diet or substance use.
We will exclude outcome data based on non‐validated tools for outcome measurement. To be classified as validated tool in this review, measurement tools (e.g. questionnaires, scales) must have been published in a peer‐reviewed journal together with standard measures of instrument quality.
Search methods for identification of studies
Electronic searches
We will conduct systematic literature searches using a broad and sensitivity‐maximising approach. We will not use study design filters as recommended for reviews in public health and health promotion (Higgins 2011). This is in order to obtain a comprehensive summary of published and unpublished evidence and to assess generalisability of findings across different implementations of the intervention. We will conduct the literature searches in English‐language databases but we will include all eligible studies regardless of the language in which the studies are published. As far as possible, we will involve native speakers for assessing inclusion criteria and data extraction of studies published in languages in which the review authors are not proficient. The literature search will include the following databases from their inception up to the most recent version:
-
CENTRAL (The Cochrane Library);
-
MEDLINE (PubMed);
-
EMBASE (embase.com);
-
PsycINFO (ProQuest) (Appendix 1);
-
bibliographic databases of the National Institute for Occupational Safety and Health (NIOSH) (NIOSHTIC and NIOSHTIC‐2);
-
the Health and Safety Executive (HSE)'s HSELINE Database;
-
the International Occupational Safety and Health Information Centre (CIS) bibliographic database, CISDOC.
Searching other resources
In order to identify ongoing and unpublished trials, we will search the following trial registers:
-
the WHO International Clinical Trials Registry Platform (ICTRP) search portal (http://apps.who.int/trialsearch/);
-
the Trials Register of Promoting Health Interventions (TRoPHI) (http://eppi.ioe.ac.uk/webdatabases/Search.aspx);
-
ClinicalTrials.gov (http://clinicaltrials.gov/).
We will check reference lists of all included studies and of relevant review articles for additional references. We will contact experts in the research field to identify unpublished data, ongoing studies or studies that have been missed by the literature searches.
Data collection and analysis
Selection of studies
Two review authors (AK, JG) will independently screen titles, keywords and abstracts of retrieved references for inclusion of potentially eligible studies. We will code references as 'retrieve' (eligible or potentially eligible/unclear) or 'do not retrieve'. We will obtain the full‐text publications of all references code as 'retrieve'. Subsequently, we will independently screen the full‐text to decide on inclusion or exclusion of studies. For excluded studies we will record the reasons in the table 'Characteristics of excluded studies'. We will resolve disagreement by consensus or third party adjudication (DN, ER or EVE). We will record the selection process in sufficient detail to be able to complete a Preferred Reporting Items for Systematic Reviews and Meta‐Analyses (PRISMA) flow diagram (Moher 2009).
Data extraction and management
Two authors (AK, JG) will independently extract data from included studies into a pre‐defined data extraction form. We will resolve any disagreements by discussion or third party adjudication where needed (DN, ER or EVE). One author (AK) will transfer data into the Cochrane Collaboration's statistical software, Review Manager 2013. A second review author (ER or EVE) will spot‐check that data is entered correctly by comparing the data entered in Review Manager 2013 with the data from extraction forms.
Assessment of risk of bias in included studies
Two authors will independently assess the risk of bias of the included studies so that one author (AK) assesses all included studies and the remaining authors (JG, ER, EVE) will assess an equal share each.
We will assess the risk of bias of the different study types as per the following methods.
-
RCTs by using the Cochrane Collaboration's tool for assessing risk of bias as described in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011).
-
cluster RCTs by using the same tool but will amend it by adding the following sources of bias particular to cluster RCTs as described in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011):
-
recruitment bias;
-
baseline imbalance;
-
loss of clusters;
-
incorrect statistical analysis;
-
comparability with individual randomised trials.
-
-
CBAs by using the risk of bias criteria developed by the Cochrane Effective Practice and Organisation of Care (EPOC) Group (Cochrane EPOC Group 2013a).
We will grade each included study in each domain of bias as either having a high, low or unclear risk of bias. We will also provide a quote from the study report to justify our judgment. When study authors do not report information that can be quoted but we can infer the necessary facts we will report our own justification for our judgment. At the study level, we will consider studies to have a high risk of bias when we judge one or more key domains to have a high risk of bias. In RCTs, we will consider random sequence generation, allocation concealment, selective outcome reporting, and incomplete outcome data to be key domains. In CBAs, we will consider if baseline outcome measurements were similar, if baseline participant characteristics were similar, if incomplete outcome data were adequately addressed, if knowledge of the allocated interventions was adequately prevented during the study, if the study adequately protected against contamination and if the study was free from selective outcome reporting to be key domains.
We will summarise and present risk of bias data in a 'Risk of bias summary figure' together with a 'Risk of bias graph' as described in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011). Where information on risk of bias relates to unpublished data or correspondence with a trialist, we will note this in the 'Risk of bias' tables.
Measures of treatment effect
We will use risk ratios (RRs) for dichotomous outcomes, and mean differences (MDs) or standardised mean differences (SMDs) for continuous outcomes, or other type of data as reported by the authors of the studies. If only effect estimates and their 95% confidence intervals (CIs) or standard errors are reported in studies we will enter these data into Review Manager 2013 using the generic inverse variance method. When the results cannot be entered in either way, we will describe them in the 'Characteristics of included studies' table, or enter the data into additional tables. As recommended by Verbeek, we will record time‐to‐event data, such as time to return to work (e.g. measurement of absenteeism) as continuous outcome measures using the means given in the study groups including standard deviations (Verbeek 2009).
Unit of analysis issues
For some cluster‐randomised trials reports may contain sufficient data to be included in a meta‐analysis but no allowance for the design effect. In these cases, we will calculate the design effect based on a fairly large assumed intra‐cluster correlation of 0.10. We consider this assumption a realistic estimate based on current best practice with studies about implementation research (Campbell 2001). We will follow the methods stated in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011).
Dealing with missing data
To complete data on intervention or outcome or to clarify methodological details, we will contact first or last authors by e‐mail or by telephone. If essential information cannot be obtained within a reasonable time after contacting the authors, we will record the respective risk of bias domain as 'unclear'. If applicable, we will assess the effect of missing data on findings in sensitivity analyses.
If numerical outcome data such as standard deviations or correlation coefficients are missing and cannot be obtained from the authors, we will calculate them from other available statistics such as P values according to the methods described in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011).
Assessment of heterogeneity
We will consider interventions as similar when they can be grouped within the same category of HRM training interventions as defined in Table 1. For example, we will consider a web‐based lecture aiming to improve active listening similar to a face‐to‐face lecture on how to improve supervisory emotional support to employees. In contrast, we will consider an individual in‐house coaching session aiming to improve interactional justice dissimilar to a role playing simulation on how to provide immaterial reward and respect to the employee.
We will consider outcome measurements as similar when they fall into the same outcome category (either measures of psycho‐mental stress, measures of absenteeism or measures of well‐being). Drawing on Naghieh 2013, we will regard follow‐up times of less than three months (short term), three months to one year (medium term) and more than one year (long term) as different.
We will use the I2 statistic and the chi2‐test to assess statistical heterogeneity in results. We will refrain from meta‐analyses if there is evidence of considerable statistical or clinical heterogeneity (I2 > 75%).
Assessment of reporting biases
We will assess reporting bias according to the recommendations given in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011). This will include appraisal of publication bias, time‐lag bias, multiple publication bias, location bias, citation bias, language bias, and outcome reporting bias. If applicable (more than five studies), we will assess publication bias by drawing a funnel plot and undertaking Begg's and Egger's tests. To avoid language bias we will include studies irrespective of language of publication. When we identify multiple reports on the same study, we will link each publication to one reference we choose to represent them all as recommended in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011). We will analyse outcome reporting bias by comparing outcome measures stated in Methods sections (or other sources such as trial registries, or study protocols) with those in study reports.
Data synthesis
If two or more studies are sufficiently homogeneous regarding the category of intervention and category of outcome measure, we will perform a meta‐analysis of outcome data. We will base our conclusions on RCTs but we will also report results from CBAs as supporting evidence. We will perform meta‐analyses separately for RCTs and non‐randomised trials. We will analyse cluster RCTs together with RCTs if they adjusted their results for the design effect. Given the expected heterogeneity amongst our included studies we will calculate pooled point estimates and 95% CIs using a random‐effects model. We will use Review Manager 2013 to generate summary statistics, conduct meta‐analyses, and produce forest plots. Where clinical or statistical heterogeneity is considered too large for conducting meta‐analysis, we will perform a narrative synthesis. We will support the narrative synthesis of studies by creating harvest plots (Ogilvie 2008; Petticrew 2013).
The harvest plot is a novel and useful aid to synthesising evidence about the effects of complex, heterogeneous, population‐level interventions (Ogilvie 2008). Drawing on methods described by Ogilvie 2008, harvest plots were developed to visually convey findings, appropriateness of the study design, confidence in the estimate of effect and risk of bias of the included studies; this method has already been successfully used in other systematic reviews of complex interventions (Crowther 2011; Thomas 2008; Turley 2013).
We will create a 'Summary of findings' table using the pre‐specified primary outcomes categories psycho‐mental stress, absenteeism, and well‐being.
We will use the Grading of Recommendations Assessment, Development and Evaluation (GRADE) approach to assess the quality of the body of evidence for a given primary outcome category, classifying the quality of evidence as high, moderate, low, or very low (GRADE Working Group). We will present the quality of evidence in one or several Summary of findings tables, using methods and recommendations described in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011). We will justify all decisions to downgrade the quality of the evidence using the five criteria of study limitations, consistency of effect, imprecision, indirectness and publication bias, or to upgrade the quality of evidence using the three criteria of strength of effect, dose‐response relationship and residual confounding. We will make comments to aid readers' understanding of the review where necessary.
Subgroup analysis and investigation of heterogeneity
If applicable, we will conduct the following subgroup analyses.
-
Company size (small and medium‐sized vs. larger (>250 annual work units) enterprises (European Commission 2003). The hierarchical structures and shared values of large enterprises may differ from that in small companies. We assume that the effect of HRM training aiming to change the attitude and behaviour of supervisors may also be influenced by higher level managers and corporate culture.
-
Economic sector (primary, secondary, tertiary (person‐oriented services or non‐personal services)). Employees' needs (e.g. security of income and employment) and work‐stressor profiles (e.g. physical stress in agriculture or construction vs. psycho‐mental stress in human health services) may differ according to the sector of economic activity.
-
Income level of the country the study was conducted in (high‐income vs. low‐ or middle‐income) (The World Bank 2013). We expect that differences in the ethical, cultural and economical background of the country the study was conducted in may influence our outcomes of interest. We assume that especially the cultural and economical background would have a large impact on employees' attitudes and needs.
-
Concept of training (theory‐based vs. not theory‐based). Theory‐based training interventions provide established psychological explanations for relationships between supervisor behavior and employee reactions or outcomes (e.g. hypothesised processes of full range leadership model or dyadic leadership) whereas interventions not based on theory are merely exploratory.
-
Type of training (Table 1).
Sensitivity analysis
If applicable, we will examine the robustness of pooled estimates and variance by excluding:
-
studies with a high risk of bias;
-
unpublished studies;
-
studies in which outcome was measured in indirectly subordinated employees (second level or higher) or in groups of both, directly and indirectly subordinated employees (as far as these information are available).
Logic model of how HRM training of supervisors may lead to positive changes in employees
| Training off‐the‐joba — Self‐help activities (textbook‐based, video‐based, web‐based) — Formal training (face‐to‐face lecture, classroom lecture), case analysis and discussion— Simulations— Role playing — Behavioural role modelling — Developmental assessment centres | Training on‐the‐joba, (real time/things/situations) — Executive Coaching — Mentoring — (360°) Feedback — Job rotations — Action learning | |
| Supervisor‐Employee Interaction — Communication (e.g. active listening, clear instructions, role clarity, clear and ethical organisational and personal goals, etc.) — Justice (interactional) — Recognition and immaterial reward (appreciation, respect, etc.) — Supervisory support (emotional) — Leadership style | (A1) e.g. web‐based teaching of how to improve active listening, role‐playing simulation for improving active listening skills | (A2) e.g. personal coaching concerning active listening while working |
| Design of working environment b — Participation — Justice (procedural and distributive, including material reward) — Autonomy (e.g. working time/schedule) — Decision latitude and control — Supervisory support (instrumental) | (B1) e.g. classroom lecture for improving knowledge on positive health effect of participative leadership, case‐analysis simulation for improving participative leadership behavior | (B2) e.g. 360°‐feedback session for assessing and improving the leaders participative leadership behavior |
| (a) For example, combined interventions containing both, an introduction to active listening skills using a video‐based lecture (A1) followed by personal coaching sessions (A2) will be classified as A2. For another example, a short video‐based introduction on active listening (A1) followed by an considerably longer role playing session on participative leadership (B1) will be classified as B1. (b) Only included if these conditions could (and should) be influenced directly by the targeted supervisor. | ||
| Work‐related stressors | Work‐related resources |
|
|
|
| The list is neither exhaustive nor does it make any claims about completeness. Some factors can be both, a work‐related stressor and a resource depending on whether they are present/pronounced or absent/less pronounced (Cropanzano 2005;Eurofound 2010; Houtman 2007;Kals 2012; Lohmann‐Haislah 2012; Semmer 2010; Stadler 2003;WHO 2013; WHO Regional Office for Europe 2010). | |
