Employee engagement and performance: a systematic literature review

Employees, regardless of the nature of business, are one of the key resources of a company. At a time when competition for the best specialists is often stronger than efforts to obtain the client, the ability to successfully manage relationships with employees can decide on the long-term market advantage. Through the creation of a friendly, development-enabling environment that leads to employee engagement, employers can increase their chances of hiring and retaining valuable employees.

Numerous reports suggest a low level of employee engagement worldwide. According to the Gallup Institute, globally only 15% of workers can be described as fully engaged in their work, while 85% are not engaged or are actively disengaged [Gallup, 2017]. The causes of the deepening “disengagement crisis” lie in the absence of support to employees in achieving what they perceive for themselves as meaningful results [Forbes, 2014]. The academic community also sees the problem of low employee engagement and its negative impact on business outcomes. According to Teresa Amabile of the Harvard Business School, cited in Forbes [2014], it leads to lower level of company revenues and deterioration of its profitability indicators. The crisis, in the context of innovation, productivity, and performance, has also been noticed by the UK Government. The Department for Business, Innovation and Skills [2009] has confirmed the low level of employee engagement in Great Britain and the negative consequences of this state of affairs on the UK economy. The worldwide nature of the problem suggests the need for research on the impact of employee engagement on the organization, as well as synthesis of the results, in an attempt to draw practical, universal lessons.

Tranfield et al. [2003] indicate that the process of reviewing the literature is essential for managing knowledge diversity in the context of particular academic studies. This article is a systematic literature review, with the objective of ordering and categorizing prior research on the association between employee engagement and various performance aspects. It identifies gaps in the current state of knowledge, draws conclusions, and highlights the implications based on analysis of selected studies. The review includes papers published between January 2002 and May 2018.

The article structure is as follows. The first section describes the method of selection of publications. Next, the analysis results are presented. The last section contains conclusions of the review, their practical and academic implications, as well as the paper’s limitations.

A review of literature lays a solid foundation for knowledge development. It exposes areas where studies are yet to be conducted and assists in theory advancement [Webster et al., 2002]. This article is of the systematic review type, with elements of bibliometric analysis. It is aimed to synthesize the current state of knowledge by presenting a structured analysis and the aggregated findings. For academics, this type of review enables the maintenance of scientific rigor. For employers/managers, the reviewing process facilitates accumulation of credible knowledge from various sources [Tranfield et al., 2003]. In carrying out the review, a comprehensive search strategy with an explicit process of selecting articles is being set out. The process of extracting relevant information and evaluating articles is clear and specific. Categorizing studies within conceptually and/or methodologically similar groups leads to systematic ordering and description of the results of the studies in order to assess their quality and applicability. Systematic reviews, apart from summarizing the current state of knowledge within an area, also clarify what remains to be examined [Petticrew and Roberts, 2006].

This article also contains elements of descriptive bibliometric analysis. They were included in order to determine the distribution of reviewed publications over the years and across different geographical regions. The applied approach offers insight on high levels of aggregation and, therefore, does not allow for more detailed conclusions [Van Leeuwen, 2004].

Only few literature reviews related to the concept of employee engagement have been conducted to date. Three of them seem to have had the most significant impact on the direction of further research. Wollard and Shuck [2011] carried out a literature review, which explored conceptually and empirically driven antecedents of employee engagement in two perspectives: individual and organizational, finding 21 antecedents in each dimension. Apart from being an important source of information for researchers, the article is especially useful for practitioners interested in increasing engagement in their organizations.

Kim et al. [2012], in their review, investigated the consequences of employee engagement. They particularly analyzed the literature on the relationship between employee engagement and performance, presenting conclusions and implications for human resource (HR) management and organizational development. The search for relevant literature was limited to ProQuest multiple databases, including Education Resources Information Center (ERIC), ABI/ INFORM Complete, ProQuest Education Journals, PsycARTICLES, and PsycINFO. Analysis of 20 articles selected for the final review proved a positive association between employee engagement and company performance, concluding that an engaged workforce can be an important source of competitive advantage [Kim et al., 2012, p. 267].

Bailey et al. [2017] conducted a systematic review, involving 214 studies that examined the meaning, antecedents, and consequences of engagement (42 studies researched the performance outcomes). They distinguished five groups of engagement’s determinants: leadership, job design, team and organizational factors, organizational interventions, and psychological states. Engagement was found to be positively associated with four work-related aspects: individual morale, individual task performance, organizational performance, and extra-role performance.

This literature review can be regarded as an extension of Kim et al.’s [2012] as well as Bailey et al.’s [2017] substantial works, in terms of publication date range and databases used in order to extract relevant articles. Thus, it continues to drive the explorative conversation around the construct of engagement in research.

The review includes only peer-reviewed articles from three databases: EBSCO (Business Searching Interface), Emerald, and ProQuest (ABI/INFORM Collection). The search was confined to English language publications that contained the expression “employee engagement” or “work engagement” in the abstract, the most frequent keywords used to describe the analyzed construct [Bakker and Bal, 2010]. At this stage, the selection yielded >2,000 publications. In the next step, the search was limited to articles with abstracts that contained the word “performance”, the second determinant of the analyzed dependency. This restriction reduced the pool to 550 articles. The selection process is consistent with the approach applied in earlier reviews of the literature on the subject [Kim et al., 2012].

Figure 1 shows the distribution of publications over time. The trend is positive, with a particular increase in the number of articles in the period 2013–2018.

Figure 1

The final step of article selection was verification of abstracts, performed to ascertain whether they present the measurement and results of the authors’ own empirical research. Finally, the literature review includes 71 publications that have been analyzed in detail.

The literature review comprised an analysis of articles that met the criterion of conducting research and describing the results thereof on the association between employee engagement and performance aspects. The papers were published in 49 different journals related mainly to the following academic disciplines: occupational psychology, HR management, and performance management. The most studies were published in the Journal of Occupational and Organizational Psychology (n=6), International Journal of Contemporary Hospitality Management (n=3), Journal of Managerial Psychology (n=3), Journal of Organizational Behavior (n=3), and Personnel Review (n=3).

The examined studies were conducted in 25 countries, across five continents (Table 1). This observation demonstrates the wide interest of various academic communities in this topic. Most of the research has been conducted in the USA (10 studies) and the Netherlands (10 studies). The sum of publications in Europe alone accounted for 46% of all reviewed papers.

Participants of the research studies represented diverse professions, ranging from manufacturing to the services sectors; 44% of the articles described the results of studies carried out on more than one professional group. Among the studies limited to representatives of one branch of the economy, most concerned the following industries: hospitality (11%) [Park et al., 2017; Kim and Koo, 2017; Karatepe and Olugbade, 2016; Suan Choo, 2016; Myrden and Kelloway, 2015; Karatepe, 2011; Xanthopoulou et al., 2009; Salanova and Agut, 2005], banking (8%) [Lathabhavan et al., 2017; Karatepe and Aga, 2016; Akhtar et al., 2016; Van Beek and Taris, 2014; Albdour and Altarawneh, 2014; Yalabik and Popaitoon, 2013], and health care (6%) [Shantz et al., 2016; Gordon and Demerouti, 2015; Tims and Bakker, 2013; Freeney and Fellenz, 2013].

Petticrew and Roberts [2006, p. 170] indicates that “systematic ordering and description of the results of the studies lies at the core of most systematic reviews”. Therefore, the main research objective of the review was to categorize the analyzed studies and draw conclusions based on their ordering. Categories were created in order to map out studies with regard to the examined performance aspects associated with employee engagement. Campbell et al. [1993] indicate a wide variety of phenomena under the “performance” label in contemporary research and enumerate the measures of performance, ranging from objective ones, like number of pieces produced or total value of sales, to subjective self-ratings of overall performance. Performance can be defined as work activity and its dynamics but also as a measurable work outcome [Roe, 1999]. Furthermore, it can be considered on individual, group, or organizational level, which is reflected in specific goals that correspond with each level [Roe, 1999]. Based on this description, six main categories, considering the type and level of performance, have been created: process performance on the individual level, process performance on the team level, process performance on the organizational level, outcome performance on the individual level, outcome performance on the team level, and outcome performance on the organizational level. Each study has been assigned to appropriate category/ categories (eight studies have been assigned to two categories, while one study has met the selection criteria of three categories). Table 2 illustrates the result of the ordering and categorization of the analyzed papers.

Categorization revealed an uneven distribution of the analyzed studies across six defined categories. The majority of the research reports (61 out of 71 papers) were assigned to process performance on the individual-level category, nine papers examined the association between employee engagement and process performance on more aggregated (team and organizational) levels, and 10 studies linked employee engagement with outcome performance categories on various levels, with one [Gorgievski and Moriano, 2014] examining both individual and organizational levels. High concentration of studies in a single category is the main characteristic of prior literature, which analyzed the association between employee engagement and performance. It shows that there is still a vast research area requiring further exploration. Future studies can contribute to the body of knowledge by focusing more on links between employee engagement and outcome performance aspects, as well as association between employee engagement and performance on aggregated levels.

The main process and outcome performance categories can be further subcategorized based on well-established definitions. Borman and Motowidlo [1997] divide process performance into contextual and task performance, whereas Lebas and Euske [2002] divide outcome performance into financial and nonfinancial performance. Ordering of the research studies based on a more detailed categorization is presented in Table 3 (the total number of studies in the subcategories is higher than the total number of papers in categories due to the fact that many studies examined more than one subcategory of performance).

Subcategorization of the main performance categories showed relatively even distribution of studies across subcategories. Process performance on individual-level category was well represented in both task and contextual performance subcategories, with at least 40 studies in each of them. Apart from the deficiency of studies related to process performance on the organizational level, future research should pay more attention to examining the links between engagement and contextual performance on the team level. It would be particularly valuable, as contextual performance is, by definition, a discretionary behavior directed toward an employee’s proximal environment, including cooperation within a team or unit [Borman and Motowidlo, 1997]. Studies researching the association between employee engagement and outcome performance, despite their general scarcity, are evenly distributed across the subcategories. The agenda of future research ought to include more studies on both financial and nonfinancial dimensions of outcome performance and their relation with work engagement, especially since those aspects of performance may be particularly meaningful and persuasive for employers.

Analysis of individual studies required verification of authors’ approach to defining the “employee engagement” concept. The reason is that construct clarity enables to avoid the hampering of the knowledge accumulation process in a particular research field [Farndale et al., 2014]. The objective of this phase was to avoid a situation where significant differences in definitions would undermine the credibility of the review.

In the literature, there are several terms that describe work-related engagement (Table 4). These include in particular, “employee engagement”, “work engagement”, “organization engagement”, and “job engagement.” The first two, according to Schaufeli and Bakker [2010, p. 10], “are typically used interchangeably” and considered so for the purpose of the review. The majority of authors (76% of the publications) adopted the definition proposed by Schaufeli et al. [2002, p. 74], which describes employee/ work engagement as “a positive, fulfilling, work-related state of mind that is characterized by vigor (e.g., being highly energetic), dedication (e.g., being highly involved in work), and absorption (e.g., being highly concentrated in work)”.

However, some authors emphasize the need for separate definitions and analyses of the two types of engagement, as the state associated with a particular workstation and the state associated with affiliation to the organization. Saks [2006, p. 604] calls the first of these concepts “job engagement”, defined as “the extent to which an individual is psychologically present in his work role”. Analogously, “organization engagement” is “the extent to which an individual is psychologically present in a role as a member of an organization”. From the employer’s point of view, it is important to quantify engagement level in both dimensions. According to Farndale et al. [2014], this allows to relate measurement results to the role-specific outcomes of engagement, which may differ at the organizational and job levels.

Despite the varying levels of details in the definitions, all of them are based on Kahn’s concept of personal engagement. Therefore, it can be assumed that the reviewed publications belong to the same conceptual approach and comparison of their results is justified. In the case of six papers [Anitha, 2014; Coco and Jamison, 2011; Dalal and Baysinger, 2012; Dash and Muthyala, 2016; Medlin and Green, 2009; Suhartanto and Brien, 2018], the authors mentioned several definitions without adopting one specific concept of employee engagement.

Consistency of definitions is also reflected in a limited number of tools used by the researchers to quantify the level of employee engagement: 82% of the analyzed publications have benefited from different versions of the Utrecht Work Engagement Scale (UWES) developed by Schaufeli and Bakker of Utrecht University [Schaufeli et al., 2002]. In its original form, the scale consisted of 25 statements associated, in accordance with the definition described earlier, with the three work-related engagement dimensions: vigor, dedication, and absorption. Study participants assign a point value to each statement using a seven-point scale, where zero means “never” and six means “always”. The most commonly applied version of the scale in the analyzed studies was UWES-9, in which three statements were assigned to each dimension. This variant was used in 48 out of 71 publications.

Among the other tools used by the researchers, the most popular were Saks Engagement Scale [Saks, 2006] and the Gallup Workplace Audit [Harter et al. 2002], both consisting of 12 items rated on a five-point Likert scale. In some cases, the authors applied a combination of several tools (Table 5).

Anitha [2014] applied a self-designed survey questionnaire, whereas Coco and Jamison [2011] did not indicate the survey that was used in their study. Dalal and Baysinger [2012] created a composite scale that included items from other tools such as UWES, Job Engagement Scale, and Work Engagement Scale.

Despite the fact that the UWES was used in the majority of the reviewed papers, there are doubts about the reliability of the questionnaire. Kulikowski [2017], in his literature review, analyzed studies on the UWES factorial validity. As the results were found to be ambiguous, he indicates the importance of routinely testing the UWES factorial validity during research on employee engagement. Furthermore, concerns are raised about the discriminant validity of UWES with regard to job satisfaction [Viljevac et al., 2012]. Acknowledging the contribution that UWES has made to the current understanding of performance aspects in the workplace, there is a necessity for further search for better, more internally consistent work engagement measures [Kulikowski, 2017].

In 43 out of the 71 analyzed studies, authors relied solely on self-reported questionnaires in order to collect source data on the work engagement outcomes. Consequently, the data used to quantify the dependent variable and the explanatory variable were provided by the same person, which could, to some extent, lead to overestimation of the correlation between those variables. Subjectivity of that method stems from the tendency of respondents to maintain consistency between the declared attitudes and actions [Podsakoff et al., 2003]. On the other hand, researchers have used proven, widely accepted questionnaires that allow a reliable evaluation of examined variables. In the remaining 28 studies, authors used more objective data, such as financial results of business units, surveys conducted among customers, assessments of direct supervisors, or a set of several aforementioned data sources (Table 6).

Various authors, in order to verify their hypotheses, have applied various techniques of statistical modeling. The most commonly used was structural equation modeling (including variants such as path analysis, partial least-squares analysis, and confirmatory factor analysis), applied in 42 reviewed studies. One of its standard applications is research in the management science using scale-based questionnaires [Davcik, 2014]. In the remaining studies, researchers have benefited from other modeling approaches, such as multilevel linear regression, Bayesian methods, or canonical discriminant analysis (Table 7).

The vast majority of analyzed studies had a cross-sectional structure of source data, thus they were conducted at one point in time. The consequence of this type of analysis design is the difficulty in determining the causal inference between variables, which, in the context of this review, is the association between employee engagement and various aspects of performance. However, it is proven that meeting certain requirements can strengthen causal inference and mitigate the risk of common-method variance [Rindfleisch et al., 2008]. According to Rindfleisch et al. [2008, p. 276], cross-sectional data are applicable especially for studies that are well embedded in theory, use reliable measurement scales and statistical tools, and involve well-educated respondents.

In the case of 13 analyzed studies, the source data had a longitudinal or time-lag structure, which means that the variables had been quantified at least twice over time [Alessandri et al., 2018; Bal and De Lange, 2015; Halbesleben and Wheeler, 2008; Karatepe, 2011; Karatepe and Aga, 2016; Karatepe and Olugbade, 2016; Khan and Malik, 2017; Myrden and Kelloway, 2015; Tims et al., 2015; Wang et al., 2015; Xanthopoulou and Bakker, 2008; Xanthopoulou et al., 2009; Yalabik and Popaitoon, 2013]. Successive measurements of the same variables enable more unambiguous identification of the causal link between them. However, a drawback is the risk of reducing the rate of responses received for repeated measurements, as well as the occurrence of intermediate events, which could disrupt the results of a study [Burkholder and Harlow, 2003].

The next division of reviewed publications can be made by taking into account the directness of relationship between employee engagement and its consequences. Among the analyzed articles, in 89% of the cases, the authors have examined only the direct relationship between the variables. In other publications, the link has been supplemented by an additional mediating factor (motivation, workplace optimism, service climate, team orientation, and learning goal orientation). Their introduction was aimed at strengthening the original relationship, which has been proved by the studies’ results.

Forty-eight studies found a statistically significant relationship between employee engagement and task performance, although one study concluded that only organizational engagement (one of the engagement dimensions) is related to the individual level of task performance [Kim and Koo, 2017]. Forty-six studies examined the association between work engagement and contextual performance in terms of the following behavioral aspects: organizational citizenship behavior, extra-role behavior, innovative behavior, employee retention (positive), turnover, absence intention (negative), organizational and career commitment, initiative, active learning behavior, knowledge sharing, creativity, proactivity, counterproductive behavior (negative), adaptability, decision-making quality, and safety behaviors. Thirty-six of these studies fully confirmed the formulated hypotheses, whereas 10 showed mixed results. Gordon and Demerouti [2015] found that work engagement was an antecedent only of daily analytical decision-making, however, not of intuitive decision-making. Halbesleben and Wheeler [2008] revealed that engagement was not significantly related to turnover intention (negative relation). Two studies found that work engagement was not associated with organizational citizenship behavior [Tims et al. 2015, Zhong et al. 2016]. Saks [2006] concluded that only organization engagement predicted organizational citizenship behavior. Albdour and Altarawneh [2014] and Farndale et al. [2014] confirmed that organization engagement was not significantly related to continuance commitment, one of the organizational commitment components. Reijseger et al. [2017] found that employee engagement was not associated with counterproductive behavior (negative relation). Kim and Koo [2017] concluded that organization engagement was not significantly related to innovative behavior. In the study of Shuck et al. [2015], the relationship between employee engagement and behavioral outcomes depended on the scale used to measure engagement. Results were statistically significant when UWES-9 was applied but inconclusive when the Job Engagement Scale was used.

Seven studies found a statistically significant link between employee engagement and financial performance (revenue, profit, and inventory shrinkage cost), although one concluded that engagement was associated only with profit but not turnover [Dijkhuizen et al., 2016]. Seven studies focused on the association between engagement and nonfinancial performance (environmental performance, customer satisfaction, number of employees, and safety level). In six of these studies, the hypotheses formulated by the authors were confirmed, whereas one did not find a statistically significant link between entrepreneur engagement and the level of company’s employment [Dijkhuizen et al., 2016].