Supporting temporary agency workers’ affective commitments: exploring the role of opportunities for competence development

ABSTRACT

Nowadays, many organizations employ parts of their workforce via employment agencies and, as such, temporary agency workers (TAWs) are important for our economies. Increasingly, research is conducted to explore the relationships among the three parties involved – the TAW, the client organization and employment agency – and how this relationship can be strengthened by HR instruments. This study adds to this stream of literature by studying to what extent TAWs’ affective commitment (AC) towards the client organization and employment agency relates to their expectations and their fulfilled expectations regarding offered opportunities for competence development (OfCD) by both organizations, and by exploring to what extent TAWs regard both employment organizations responsible for offering them OfCD. A cross-sectional survey study was conducted among TAWs in the Netherlands (n = 449), including quantitative and qualitative analyses. All results of this study together revealed that TAWs tend to be focused on the client organization with regard to OfCD. This is the organization to which they feel most affectively committed, that they regard responsible for organizing their OfCD, and where they ask for OfCD. This study indicated that TAWs do not yet profit maximally from the triangular employment relationship, and have a rather traditional view on TAW constructions.

KEYWORDS:

• Temporary agency workers
• opportunities for competence development
• affective commitment
• client organization
• employment agency

Introduction

During the past two to three decades, it has become common practice for organizations in Western economies to seek alternatives for ‘standard’ contracts (i.e. full-time, indefinite period of time) as a means to maintain staff flexibility and at the same time maintain a competitive position (Chambel et al. 2015; Connelly and Gallagher 2004). One of these ‘nonstandard’ or ‘atypical’ employment contracts includes temporary agency work: an employment construction in which employees are officially employed via an employment agency and work for a client organization for a limited period of time (Liden et al. 2003). Temporary agency workers (TAWs) are thus employed by means of a triangular employment relationship as there is an employment relationship between the TAW and the employment agency, a management relationship between the TAW and his/her client organization, and a business relationship between the client organization and the employment agency (Håkansson and Isidorsson 2015). During the past 10 years, the number of TAWs remained quite stable and accounted for 1.8% of the total employment in 2012 (Eurociett UNI Europa 2015). Despite the fact that the percentage of TAWs as part of the total European employment is small, TAWs play a key role in the labour market and are of high importance for all three parties involved (Eurociett UNI Europa 2015; Schmidt and Thommes 2007).

Theoretically, it is advantageous for both client organizations and TAWs to use TAW constructions as these facilitate optimal distribution of flexibility and security (flexicurity paradigm, cf. Wilthagen and Tros 2004) since three rather than two parties are involved (Dekker and Wilthagen 2014). In the short term, TAW constructions decrease unemployment for employees. In other words, the flexibility enables organizations to create jobs that would otherwise not exist (Eurociett UNI Europa 2015). Moreover, in the long term, it allows them to improve their employability by trying out various jobs/roles in different client organizations (e.g. Felfe et al. 2008; Van Breugel, Van Olffen, and Olie 2005). From the perspective of ‘boundaryless careers’ (e.g. Bravo et al. 2015), it is argued that employees increasingly pursue careers in which they are flexible and free to occasionally switch between jobs and roles. TAWs can therefore satisfy their individual needs and preferences, such as freedom and flexibility (cf. Felfe et al. 2008), knowing that the employment agency will guarantee job security. For client organizations, it reduces costs and increases flexibility (e.g. Van Breugel, Van Olffen, and Olie 2005), as they can attract highly educated employees for jobs/roles requiring specific knowledge or skills, or employees for jobs/roles that are relatively easy to learn and to perform, and end the contract when the task is fulfilled. Working with TAWs therefore ensures that maximum benefit is gained from new blood, new knowledge and new skills and the external view TAWs bring to the organization (Gallagher and McLean Parks 2001).

However, the extent to which these theoretical notions on the benefits of TAW constructions hold true in practice is questionable. More specifically, a triangular employment relationship may lead to more complexity with regard to the ‘psychological contract’ between the parties involved. A psychological contract refers to the mutual beliefs, perceptions and informal obligations between an employer and an employee (Rousseau 1995). Psychological contract theory is based on the social exchange theory (Blau 1964), which states that people tend to reciprocate favours they receive from others; if employees experience favours from their employers, they will develop a psychological obligation to display higher levels of organizational commitment and performance. Since mutual obligations among the parties involved often remain implicit, differences in interpretations of these obligations can easily occur (cf. Chambel and Castanheira 2012; Lapalme, Simard, and Tremblay 2011), with ‘contract breaches’ as a result. TAWs are likely to perceive themselves as having two psychological contracts (i.e. with each of their employment parties, Claes 2005), which doubles the risk of a discrepancy between expectations or between interpretations of mutual agreements which, in turn, may negatively influence the relationships among the three parties (Lapalme, Simard, and Tremblay 2011). Specifically, contract breaches may have substantial negative consequences for the organizational commitment TAWs experience (cf. Dekker and Wilthagen 2014; Gallagher and McLean Parks 2001) – referring to ‘a bond or linking of the individual to the organization’ (Mathieu and Zajac 1990, 171) – and it may even result in a decrease in performance and job satisfaction due to negative feelings regarding their employer(s) (cf. Meyer et al. 2002).

In this study, we examine the manner and degree to which TAWs’ organizational commitment can be explained by their perceptions of the offered opportunities for competence development (OfCD). In doing so, we make two important contributions to Human Resource Development (HRD) theory. Firstly, this study adds to current research on explaining organizational commitment, by focusing on the specific situation of TAWs. Organizational commitment is seen as a main factor that contributes to a triangular relationship in TAW constructions (cf. Dekker and Wilthagen 2014), but research on TAWs’ organizational commitment is lacking (cf. Watkins and Marsick 2014). The present study examines TAWs’ affective organizational commitment (further referred to as ‘affective commitment’, or AC), which refers to ‘the emotional attachment to, identification with, and involvement in an organization’ (Meyer et al. 2002, 21) and is generally viewed as an important predictor of employees’ performance (Chambel et al. 2015; Meyer et al. 2002).

Secondly, this study examines the extent to which OfCD, being an HR instrument, is perceived as an important return on the TAWs’ investment in being involved in a temporary employment situation. Prior to the increase in flexibility, formal employment relationships were implicitly underpinned by a psychological contract that ensured lifetime employment as a reward for employee loyalty and trust (Kornelakis 2014). However, the fact that lifetime employment in organizations is increasingly being replaced by lateral moves across jobs and organizations (Murphy and Garavan 2009) calls for new kinds of psychological contracts in which organizations contribute towards the employability of flexible employees (and specifically TAWs) in return for good performance. Since OfCD contributes, theoretically, towards TAW employability (cf. Dekker and Wilthagen 2014; Veld, Semeijn, and Van Vuuren 2015), it helps create a focus on the long term rather than the short term because of the often limited contracts. Moreover, offering TAWs OfCD contributes towards their feeling of being equally treated compared with permanent employees, which results in a positive workplace atmosphere and in turn also contributes towards the organizational commitment of permanent employees (cf. Connelly, Gallagher, and Wilkin 2014). Based on these aspects, it was expected that OfCD would be perceived by TAWs as a favour – a return on their investment – in response to which they would develop higher levels of AC. As such, following Håkansson and Isidorsson (2015), it is expected that OfCD is part of the employment relationship TAWs have with their employment agency, as well as part of the management relationship TAWs have with their client organization.

These expectations resulted in two hypotheses tested in the current study. In the following is the first hypothesis, in line with the conclusions of Lapalme, Simard, and Tremblay (2011) on the parallel character of the two separate psychological contracts that TAWs have with their employment organizations:

Hypothesis 1: The higher the TAWs’ expectations and fulfilled expectations of OfCD offered by their client organization/employment agency, the higher their AC towards this client organization/employment agency.

Inherent to the triangular character of the TAW construction, a subsequent exploration was carried out on the cross-effects of expectations and fulfilled expectations of OfCD by the one organization, on TAWs’ level of AC towards the other organization. Ideally, employment agencies serve as the linking pin between client organizations and TAWs, by dealing with the TAW’s administration and working conditions (cf. Van Breugel, Van Olffen, and Olie 2005). If TAWs feel their employment agency is taking care of them in terms of OfCD, they may, on the one hand, develop a psychological obligation to display higher levels of AC towards the client organization as this will ultimately benefit their employment agency. On the other hand, when TAWs experience favours from their client organization in terms of OfCD, this may enhance their AC towards their employment agency, since it facilitated the opportunity to work at the client organization in the first place. As such, the second hypothesis reads as follows:

Hypothesis 2: The higher the TAWs’ expectations and fulfilled expectations of OfCD offered by the client organization and vice versa, the higher their AC towards the employment agency and vice versa.

The extent to which TAWs regard their employment parties and themselves as responsible for offering them OfCD was also explored, as this information might deepen our understanding of TAWs’ expectations regarding OfCD. As such, this study aims to examine if TAWs’ (fulfilled) expectations of OfCD and their AC towards both of their employment parties are related and clarifying these results.

To this end, the following exploratory research question was formulated:

Exploratory RQ To what extent do TAWs regard their client organization and/or their employment agency as responsible for offering OfCD?

Figure 1 visually presents the triangular relationship of a TAW construction (cf. Håkansson and Isidorsson 2015), provided with the core concepts being studied: AC and OfCD.

Figure 1. Triangular relationship of a TAW construction (cf. Håkansson and Isidorsson 2015) illustrated with the concepts being studied: affective commitment (AC) and opportunities for competence development (OfCD).

Note: The business relationship was not part of this study.

The Chambel and Castanheira (2012) study was used as starting point for the present study. Chambel and Castenheira found that training had a positive influence on TAWs’ AC, and we wanted to make some additions to their empirical work. Firstly, in our study, competence development refers to the process of acquiring competences by means of formal and informal learning activities (cf. Tynjälä 2008), not just training. Formal learning refers to certificated learning activities, such as training sessions, workshops and educational programmes. Informal learning, however, refers to uncertificated learning activities, which can occur in the form of implicit learning (i.e. the capacity to learn without awareness of what has been learned), reactive learning (i.e. near-spontaneous reflection on experiences where there is little time to reflect extensively, resulting in noticing the effects of actions and recognizing future learning activities) or deliberate informal learning (such as reflecting on one’s work, sharing knowledge with colleagues, or asking for feedback from colleagues, Eraut 2004). This study’s focus on expectations and fulfilled expectations of OfCD implies a focus on planned and conscious learning activities. As such, this study explicitly concentrates on deliberate informal learning activities, leaving aside the more implicit and reactive informal learning activities.

Secondly, the study expands upon the Chambel and Castenheira study (2012) by focusing on TAWs’ perceptions instead of actual offered opportunities. We believe that it is primarily the level of perceived OfCD that influences TAWs’ AC, as this level aligns with the core of psychological contract theory, as it is about one’s perceptions of implicit and explicit mutual promises rather than the written contract and written policies (cf. Rousseau 1995; Cassar and Briner 2011). In other words, it is this level of perception that makes the triangular relationship of a TAW construction more complex than a two-way employment relationship in which only one employer is involved.

Previous studies on the influence of expected and fulfilled (or breached) psychological contracts on employees’ AC in general show that AC is likely to decrease when employees experience violation resulting from a breach (Cassar and Briner 2011). Obuya and Rugimbana (2014), however, argue that expectations need to be exceeded to have a positive influence on employees’ AC. If expectations are simply fulfilled as expected, they will be ‘taken for granted’ and will not have any significant impact. As such, the results of these previous studies indicate a relationship between OfCD and AC, but do not provide insights in the mechanisms between OfCD and AC specifically, let alone TAWs.

Thirdly, this study builds on the Chambel and Castenheira study (2012) in that we include the relationship between TAWs’ perceptions of OfCD and AC offered by both of their employment parties. Inherent to the triangular character of the TAW construction, Lapalme, Simard, and Tremblay (2011) confirmed that TAWs are involved in two different social exchange relationships, resulting in two different psychological contracts that may evolve parallel to each other.

The setting

The present study was conducted among TAWs in the Netherlands, where the number of TAWs covered 251,000 employees in 2015 (i.e. 14% of all flexible employees in the Netherlands and 3% of the total Dutch working population). In 2013, the Netherlands was one of the seven European countries (i.e. Austria, Belgium, France, Italy, Luxembourg, Netherlands and Spain) that had established ‘bipartite funds’ to stimulate the training of TAWs (Joint Eurociett/UNI Europa 2013). This investment shows that the importance of TAW competence development, and offering them enough opportunities for development, is recognized at European level.

Method

This study included a survey research, in which both quantitative and qualitative data have been collected. Specifically, this study included an explanatory and a subsequent exploratory part, in which the qualitative analysis (answering the exploratory research question) aimed to clarify the quantitative data (testing hypotheses 1 and 2).

Subjects

In total, 6709 TAWs were invited to participate in a survey study. All of them were employed via the same employment agency, and worked in a variety of client organizations. A total of 449 TAWs completed a survey (6.7% response rate). The participating TAWs (33% male) all worked in service organizations in different sectors. Their ages ranged from 18 to 70 years (M = 41 years, SD = 14 years). The participants’ level of education ranged from primary or pre-secondary vocational education to PhD, but most participants (64.3%) had completed an applied university programme, university programme or PhD. Furthermore, participants’ contracts with the employment agency ranged from zero to eight hours a week, to a full-time contract (Mode = 33–40 h a week), and participants’ employment at the client organization ranged from less than one up to three months, to more than one year (Mode = more than one year).

Instruments

For both the quantitative and the qualitative parts of this study, data was gathered by means of one survey. This survey comprised both validated scales and self-constructed questions. Open questions were also added allowing TAWs to motivate their answers.

Affective commitment

To measure AC, the Dutch version of the affective commitment scale of the Three-Component-Model by Allen and Meyer (1990) was used, a model developed by Jak and Evers (2010). It included five statements, to be scored on a 5-point Likert scale (ranging from 1 to 5, including ‘this statement applies to me: (1) ‘very much’, (2) ‘much’, (3) ‘moderately’, (4) ‘a little’ and (5) ‘a very little’’. Participants were asked to respond to these five statements first with regard to their client organization, and then with regard to the employment agency. For example, if the employment agency’s statement was ‘I experience problems my employment agency’s problems as my own’, the client organization statement was ‘I experience my client organization’s problems as my own’. Both AC scales showed good reliability (AC_{client organization}: α = .85; AC_{employment agency}: α = .93).

Expectations of opportunities for competence development

Each respondent was asked the following question individually: ‘I expected my current client organization/employment agency …’ (a) ‘… to offer me OfCD’ (b), ‘… not to offer me OfCD’ or (c) ‘I had no expectations’.

Fulfilled expectations of opportunities for competence development

Each organization was asked the following question individually: ‘To what extent did the client organization/employment agency meet your expectations regarding offered OfCD?’ ‘My client organization/employment agency offered me (a) ‘fewer opportunities than I expected beforehand’, (b) ‘… the same opportunities as I expected beforehand’, or (c) ‘… more opportunities than I expected beforehand’.

Responsibility for offering opportunities for competence development

By means of an open question, participants could indicate who they considered responsible for offering them OfCD: either the client organization, or the employment agency, or both of them. This question was formulated explicitly as an open question to give participants the opportunity to explain their statement.

Procedure

The survey was distributed using Qualtrics, an online software program. Before the actual data gathering, two pilot studies were conducted among TAWs of the participating employment agency. The first pilot study was conducted to check the comprehensibility of the survey. TAWs from one client organization completed the survey (n = 9) during a one-hour face-to-face meeting in which the TAWs were asked to complete the questionnaire and immediately reflect on it aloud. Based on this pilot study, some questions in the survey were tightened and aligned with the jargon that was usual for TAWs working for the employment agency in question. For some questions (i.e. questions asking for participants’ background information), answer categories were added or omitted. A second pilot study was subsequently conducted among TAWs from different client organizations (n = 83) to check if the scales included in the survey could be used for the specific population of TAWs instead of permanent employees. No significant changes were made to the survey as reliability and validity checks showed acceptable results.

Analyses

Testing hypotheses 1 and 2

To test hypotheses 1 and 2, quantitative analyses were conducted. Prior to these analyses, the items of AC were recoded, to align a high score of the scale with a high score of AC. Furthermore, the items on TAWs’ expectations and fulfilment were recoded into dummy variables, so that multiple regression analyses could be carried out (see later in this section for more information).

An exploratory factor analysis (EFA) was conducted to measure if AC towards the client organization and AC towards the employment agency were indeed different constructs. This EFA had an exploratory character; in other words, no restrictions were set for the number of factors. Moreover, an oblique rotation was used, since a relationship was expected between AC towards the two organizations, given the triangular character of the employment relationship (e.g. Liden et al. 2003). Table 1 presents the factor loadings of all items and shows that items 1–5 loaded on the factor representing the client organization, whereas items 6–10 loaded on the factor representing the employment agency. These two factors resulted in 63.71% total explained variance. As such, the EFA confirmed that AC towards the client organization and AC towards the employment agency are two different constructs.

Table 1. Oblique rotated factor matrix of items measuring TAWs’ affective commitment towards the client organization and towards the employment agency.

	Factor
	Client organization	Employment agency
1. I feel as if this client organization’s problems are my own	0.523	−.015
2. I feel a strong sense of belonging to this client organization	0.801	−.015
3. I feel ‘emotionally attached’ to this client organization	0.782	−.025
4. I feel like ‘part of the family’ at this client organization	0.742	−.029
5. This client organization means a lot to me	0.807	0.038
6. I feel as if this employment agency’s problems are my own	0.015	0.731
7. I feel a strong sense of belonging to this employment agency	−.038	0.903
8. I feel ‘emotionally attached’ to this employment agency	0.016	0.866
9. I feel like ‘part of the family’ at this employment agency	−.010	0.892
10. This employment agency means a lot to me	−.023	0.861

N = 449.

Multiple regression analyses were conducted to measure the influence of TAWs’ expectations and the fulfilment of offered OfCD (independent variables) on the level of AC towards the client organization and towards the employment agency (dependent variables). Firstly, analyses were run for each organization separately (hypothesis 1). Specifically, three models were run: in model 1, only control variables age and level of education were included. In model 2, the expectations of offered OfCD as predictors for the level of AC towards the corresponding organization were added as predictors to the regression model, and in model 3, the fulfilled expectations were also added as predictors to the regression model. Secondly, we measured the extent to which expected and fulfilled OfCD offered by one organization influenced the level of AC towards the other organization (hypothesis 2). Again, the same models were run, presented as models a, b and c.

A Bonferroni correction was applied (Tabachnick and Fidell 2007) because of the increased risk of Type I error as a consequence of the different tested multiple regression models. An alpha level of 2.5% was thus specified.

Answering the exploratory research question

To answer the exploratory research question, the answers to the open question were analysed qualitatively. The data was analysed bottom up. Specifically, codes were provided indicating divisions of responsibilities, resulting in different categories. This categorization was done by means of open, axial and selective coding (cf. Boeije 2010) by the first author, and the procedure was checked by, and negotiated with, the second author of this manuscript. According to Boeije (2010), open coding refers to the process of segmenting the data into meaningful fragments, coding these meaningful pieces and comparing these pieces with each other. Without restrictions, each fragment was labelled with a code that covered the content of the fragment. In this study, the complete answer of a participant was used as a starting point for analysis. The stage of open coding resulted in a total of 13 codes covering different divisions of responsibilities regarding TAWs’ OfCD. Next, the stage of axial coding refers to the process of making connections between categories or splitting them, and checking to what extent each code is unique (Boeije 2010). In this study, the axial coding stage resulted in 10 categories: four categories were combined into one category as the nuances between them appeared to be too small to distinguish them. Then, the stage of selective coding includes the process of looking for an overall structure of the codes (Boeije 2010). In this study, this stage resulted in a final set of six categories distinguishing different divisions of responsibilities regarding TAWs’ OfCD: again, two categories were integrated with another category, and two other categories were deleted after extensive discussion as these categories did not align with the overall structure. The few answers in these categories were interpreted again and could be replaced into another category.

Results

Expected and fulfilled OfCD and TAWs’ AC

Descriptive statistics

Table 2 presents the zero-order correlations between the independent and dependent variables. Mean values (SD) are also presented for the independent variables. The mean values of AC showed that, on average, TAWs felt more affectively committed towards their client organization compared to their employment agency, and there appeared to be a significant positive correlation between TAWs’ AC towards the client organization and towards the employment agency. Table 2 shows primarily positive significant relations between the independent variables, and between the dependent and independent variables. As an exception, TAWs’ expectations of OfCD from the client organization negatively related to their AC towards the client organization, meaning that the less OfCD they expected to get at the client organization, the higher their AC at the client organization or the other way around. Together, the zero-order correlations primarily show indications for parallel relationships between expectations and fulfilled expectations of OfCD and AC towards the corresponding organization.

Table 2. Zero-order correlations (Spearman) between TAWs’ expectations of OfCD, their fulfilment of OfCD, and their affective commitment towards both employment organizations.

	1	2	3	4	5	6	Mean	SD
Dependent variables
1. AC towards the client organization							3.47	0.82
2. AC towards the employment agency	0.37**						2.32	0.81
Independent variables^a
3. Expectations of OfCD at the client organization	−.20**	−.04
4. Expectations of OfCD at the employment agency	−.05	−.04	.27**
5. Fulfilled expectations of OfCD at the client organization	.20**	.10**	.19**	.09
6. Fulfilled expectations of OfCD at the employment agency	.05	.21**	.07	.28**	.19**

N = 449.

^aGiven the measurement level of the independent variables, descriptive statistics are presented in terms of frequencies. See Table 4.

*p < .05, two tailed, **p < .01, two tailed.

Table 3. Percentages of participants’ answers with regard to their expectations and fulfilled expectations of offered opportunities for competence development (OfCD), split for both employment organizations.

	Client organization	Employment agency
Expectations
Yes, I expected OfCD	45.2%	21.4%
I did not have any expectations about getting OfCD	42.8%	61.5%
No, I expected not to get OfCD	12.0%	17.1%
Fulfilled expectations
I got less OfCD than I expected	24.3%	18.9%
I got the same OfCD as I expected	56.6%	51.0%
I got more OfCD than I expected	19.2%	30.1%

N = 449.

Table 4. Results of the multiple regression analyses (Betas) measuring the influence of (fulfilled) expectations of OfCD on affective commitment towards the corresponding organization.

		Client organization^†			Employment agency^‡
Model		1	2	3	1	2	3
Expectations
1.	Yes, I expected OfCD		.22**	.26**		.11*	.21**
2.	No, I did not expect OfCD		−.07	−.04		−.04	−.06
Fulfilled expectations
a.	I got less OfCD than I expected			−.09			−.15*
b.	I got more OfCD than I expected			.18**			.24**
Control variables
i.	Age	−.04	−.02	−.01	.08	.09	.05
ii.	Level of education
	Lower secondary vocational education	.09	.10	.09	.05	.04	.07
	Senior general secondary education, pre-university education	−.06	−.05	−.07	−.11	−.09	−.08
	Senior secondary vocational education and training	.10	.11	.10	−.06	−.05	−.03
	Higher professional education, bachelor in academic higher education	−.01	.00	−.01	−.23	−.20	−.20
	Master in academic higher education, PhD	−.03	−.04	.04	−.32**	−.30**	−.30**
R^2		.03	.09	.14	.08	.10	.16
R^2 ∆			.06	.05		.02	.08

N = 449. Model 1: only the control variables were included in the model as predictors of AC. Model 2: The expectations of offered OfCD were added to the regression model as predictors of AC. Model 3: Fulfilled expectations were added to the regression model.

^†Overall model results: Model 1: R² = .03, F(6,442) = 2.073, p = .055; Model 2: R² = .09, F(8,440) = 5.304, p < .001; Model 3: R² = .14, F(10,438) = 6.857, p < .001.

^‡Overall model results: Model 1: R² = .08, F(6,442) = 6.721, p < .001; Model 2: R² = .10, F(8,440) = 5.959, p < .001; Model 3: R² = .16, F(10,438) = 9.648, p < .001.

*p ≤ .025, two tailed, **p ≤ .01, two tailed.

In addition to Table 2, Table 3 presents the distributions of the answers (percentages) for the participating TAWs regarding their expectations and fulfilled expectations of offered OfCD. Table 3 shows three aspects in particular: (1) the client organization is seen by most participants as the organization that should have offered them OfCD; (2) for both organizations, it appeared that, relatively speaking, many people did not know what to expect regarding OfCD, and this indistinctness was even larger for the employment agency as compared to the client organization and (3) compared to the client organization, relatively speaking, many TAWs were surprised by the employment agency with regard to OfCD, by answering that they got more opportunities to develop their competences at the employment agency than they initially expected.

Expected and fulfilled OfCD and AC towards the corresponding organization

Table 4 shows the results of the multiple regression analyses measuring the relationship between TAWs’ expectations and fulfilled expectations of OfCD, and their AC towards the corresponding organization (hypothesis 1).

TAWs’ AC towards the client organization appeared to be significantly associated with their expectations of OfCD from the client organization (model 2: β = .22, p < .001; model 3: β = .26, p < .001). The TAWs’ AC towards the client organization also appeared to be significantly associated with getting more OfCD from the client organization than initially expected (model 3: β = .18, p < .001).¹ In all three models, the control variables ‘age’ and ‘level of education’ showed no significant associations with TAWs’ AC towards the client organization.

TAWs’ AC towards the employment agency also appeared to be significantly associated with their expectation of getting OfCD from the employment agency (model 2: β = .11, p = .026; model 3: β = .21, p < .001). The TAWs’ AC towards the employment agency appeared to be significantly associated with getting more OfCD than was initially expected from the employment agency (model 3: β = .24, p < .001). Moreover, less OfCD from the employment agency than was expected had a significantly negative effect on TAWs’ AC towards the employment agency (model 3: β = -.15, p = .003)². As such, for both organizations hypothesis 1 was accepted: the higher the TAWs’ expectations and fulfilled expectations of OfCD, the higher their AC towards the corresponding organization.

In all three models, there appeared to be a significant negative relationship between the TAWs’ academic educational level and their AC towards the employment agency (model 1: β = −.32, p < .001; model 2: β = −.30, p = .001; model 3: β = −.30, p = .001).

Expected and fulfilled OfCD of the one organization and AC towards the other organization

Table 5 shows the results of the multiple regression analyses measuring the relationship between TAWs’ expectations and fulfilled OfCD for the one organization, and their AC towards the other (hypothesis 2). TAWs’ AC towards the client organization appeared to be significantly associated with their expectation of getting OfCD at the employment agency, although this significant result only appeared in model c (model c: β = .13, p = .018). Furthermore, TAWs’ AC towards the client organization appeared to be significantly associated with getting more OfCD than initially expected from the employment agency (model c: β = .16, p = .002)³. Control variables ‘age’ and ‘level of education’ showed no significant associations with TAWs’ AC towards the client organization.

Table 5. Results of the multiple regression analyses (Betas) measuring the influence of (fulfilled) expectations of OfCD on affective commitment towards the other organization.

AC towards the… →		Client organization^†			Employment agency^‡
Model ↓		a	b	c	a	b	c
Expectations
1.	Yes, I expected OfCD		.10	.13*		.08	.10
2.	No, I did not expect OfCD		−.08	−.09		−.09	−.08
Fulfilled expectations
a.	I got less OfCD than I expected			.02			−.03
b.	I got more OfCD than I expected			.16*			.11*
Control variables
i.	Age	−.04	−.04	−.06	.08	.08	.09
ii.	Level of education
	Lower secondary vocational education	.09	.09	.11	.05	.05	.04
	Senior general secondary education, pre-university education	−.06	−.04	−.03	−.11	−.11	−.12
	Senior secondary vocational education and training	.10	.12	.14	−.06	−.07	−.07
	Higher professional education, bachelor in academic higher education	−.01	.02	.00	−.23	−.23	−.24
	Master in academic higher education, PhD	−.03	.00	.00	−.32**	−.33**	−.32**
R^2		.03	.05	.07	.08	.10	.12
R^2 ∆			.02	.02		.02	.02

N = 449. Model a: only the control variables were included in the model as predictors of AC. Model b: The expectations of offered OfCD added to the regression model as predictors of AC. Model c: Fulfilled expectations were added to the regression model.

^†Overall model results: Model a: R² = .03, F(6,442) = 2.073, p = .055; Model b: R² = .05, F(8,440) = 2.737, p = .006; Model c: R² = .07, F(10,438) = 3.204, p = .001.

^‡Overall model results: Model a: R² = .08, F(6,442) = 6.721, p < .001; Model b: R² = .10, F(8,440) = 6.252, p < .001; Model c: R² = .12, F(10,438) = 5.771, p < .001.

*p ≤ .025, two tailed, **p ≤ .01, two tailed.

TAWs’ AC towards the employment agency appeared to be significantly associated with their fulfilled expectation of getting more OfCD from the client organization than initially expected (model c: β = .11, p = .021). Furthermore, people with an academic level of education have lower AC towards the employment agency than people with the lowest level of education (reference available. Model a: β = −.32, p < .001; model b: β = −.33, p < .001; model c: β = −.32, p < .001)⁴. Given these results, hypothesis 2 was accepted.

Organizations’ responsibilities for offering OfCD

Subsequently, TAWs’ perceptions about which organization should facilitate them with OfCD were explored, in order to deepen our understanding about the relationships between TAWs’ expectations of OfCD and their AC towards both of their employment parties (exploratory research question). Table 6 shows the distributions of TAWs’ different perceptions. It appeared that TAWs have diverse opinions about this issue. Eventually, six different categories of shared and individual responsibilities were distinguished, showing that the participants’ answers were more nuanced than just stating that either the client organization, or the employment agency, or both of them are responsible for offering them OfCD. Most participants (44%) assigned the responsibility for offering OfCD to their client organization. One of the main arguments that TAWs gave for this answer was that their client organization had a better picture of their work progress and their performance in practice compared to the employment agency, as they were physically present at the client organization most of their time. Participants also argued that the client organization was better able to organize tailor-made CD activities. They also believed that they ‘contributed’ most to the client organization, giving the client organization a certain responsibility to offer them something valuable ‘in return’.

Table 6. Division of responsibilities in offering opportunities for competence development.

Who is responsible for offering opportunities for competence development?	Percentage	Example quote
It is primarily the task of the client organization.	44	‘I have a two-year contract at this client organization, so I believe the client organization should invest in my capacities’ (participant 28)
It is primarily the task of the employment agency.	12	‘Officially, I work for the employment agency, so it is their responsibility. I do have another employer [i.e. the client organization], but this employer could be temporary. I hope to work for a longer period for [the employment agency] (participant 400)
It is the shared responsibility of both employment organizations.	29	‘Both organizations benefit from my efforts, so it is positive for both of them if I develop myself and feel well in my job’ (participant 300)
It is the shared responsibility of the client organization, the employment agency and the TAW him/herself.	6	‘An employer should take partial responsibility for your [i.e. the employee] development. The rest is your own responsibility and depends on what you want to invest in yourself’ (participant 227)
It is the TAW’s own responsibility.	4	‘It is mainly my own responsibility to see opportunities, or to ask about the possibilities for development’(participant 209)
‘I do not see any OfCD’.	5	‘None of the them [i.e. the employment organizations] have offered me opportunities to follow a course’ (participant 322)

N = 449.

Twelve per cent of the participants assigned the responsibility for offering OfCD primarily to the employment agency. Most of the TAWs argued that the employment agency is their official employer, which gives them the responsibility to offer TAWs OfCD. Participants also argued that the employment agency benefits from well-educated and well-qualified employees as this helps them place an employee in a new job after having finished another. A group of 29% considers that the employment agency and the client organization have a shared responsibility to seek OfCD. The main argument for this shared responsibility is that both organizations benefit from a well-educated employee (instead of the employment agency alone).

Apart from these three categories, two other categories were distinguished, referring to TAWs’ own role in looking for OfCD. A minority of 6% of the participants stated that looking for OfCD is the task of all parties involved, including TAWs themselves. Another 4% argued that this is primarily their own task, based on the argument that employees are responsible for organizing and building their own career. Lastly, 5% of the participants said they do not see any OfCD at all. Despite the fact that this category does not refer to a certain responsibility, we believe it is important to take this group into account as a serious signal of the difficult and sometimes unclear situation in which they are involved.

Discussion

Affective organizational commitment (AC) is perceived as the main factor that contributes to a triangular relationship in TAW constructions (cf. Dekker and Wilthagen 2014). The present study aimed to expand upon studies that investigated how TAWs’ AC can be explained. This aim was based on the premise that TAWs are a particular group of employees that are crucial for many organizations, and for our economies, although they have their own HR dilemmas and, as such, deserve particular attention in HRD research. Specifically, this study (1) examined the role of expectations and fulfilled expectations of opportunities for competence development (OfCD) on TAWs’ affective commitment (AC) towards the employment agency and towards their client organization (hypotheses 1 and 2), and (2) explored the extent to which TAWs regard their client organization and/or their employment agency as responsible for offering OfCD (exploratory research question).

With regard to hypothesis 1, TAWs’ expectations of getting OfCD from the client organization were positively related to their AC towards the client organization. Our results also indicated that eventually getting more OfCD than was initially expected contributed positively to TAWs’ AC towards the client organization, which is in line with Obuya and Rugimbana (2014). Regarding the employment agency, comparable effects were found, although the positive effects of TAWs’ expectations of OfCD from the employment agency on their AC towards the employment agency were smaller. Instead, the effect of exceeded expectations from the employment agency was larger: if TAWs’ expectations of OfCD from the employment agency were exceeded, this related positively to their AC towards the employment agency. Moreover, getting less OfCD from the employment agency than expected beforehand was negatively related to TAWs’ AC towards the employment agency, whereas this effect was not found for the client organization. It seems the TAWs’ AC towards the employment agency was more easily affected by disappointment with the offered OfCD from the employment agency, as compared to AC towards the client organization. These different effects nevertheless confirmed Lapalme, Simard, and Tremblay (2011)’s conclusion that parallel relationships exist between expected and fulfilled OfCD, and TAWs’ AC towards the corresponding organization. Therefore, hypothesis 1 was accepted.

With regard to hypothesis 2, our results showed smaller but positive effects of TAWs’ expectations of OfCD from the one organization on their AC towards the other organization. As such, hypothesis 2 was also accepted. The more the TAWs expected to get OfCD from the employment agency, the higher their AC towards the client organization. Besides, the more OfCD TAWs got from the employment agency, the higher their AC towards the client organization. With regard to AC towards the employment agency, a significant small but positive effect was found of TAWs’ exceeded expectations of OfCD from the client organization. Although the cross-effects were rather small in both models, they are interesting in view of the triangular relationship. It implies that when TAWs have the expectation that their employment agency will take good care of them, their AC towards the client organization also increases and vice versa.

With these findings, this study underscores the call for research on psychological contracts for flexible employees and for TAWs specifically. ‘Traditional’ psychological contracts, that is, in which employee loyalty and trust in organizations are reciprocated by life time employment, no longer apply for the majority of the workforce nowadays. The main challenge for HRD practitioners and HRD researchers is to find out how organizations can contribute towards TAWs’ employability in return for their performances. Instead of viewing the alignment of HRD practice with organizational performance as a starting point, HRD professionals should focus on employee benefits as a means to reconcile organizations’ demands for flexibility of on the one hand and employee security needs on the other (Kornelakis 2014). This study provides clues for HRD practitioners to how they can achieve this by means of offering TAWs opportunities for competence development. Concrete examples are provided in the scientific and practical implications section.

In addition to the relationship between OfCD and AC, this study showed significant negative relations between the academic level of education and TAWs’ AC towards the employment agency – both in the parallel model and in the cross-relations model. No significant relationship was found between the academic level of education and AC towards the client organization. These results are comparable with the results of Morf, Arnhold, and Staffelbach (2014). Also in their study, a significant negative relationship was found between highly educated TAWs and their commitment towards the employment agency, whereas this effect was not the case towards the client organization. It could be that highly educated employees do not want to commit to the employment agency as they see their involvement in the triangular relationship as a short-term solution and feel confident in finding another job – a temporary job without interference from the employment agency, or even a permanent job – given their well-developed skills and knowledge.

With regard to the exploratory research question, this study showed that TAWs have diverse opinions with regard to the responsibility for offering OfCD. The majority of them (44%) regarded their client organization as solely responsible for offering them opportunities to develop their competences. Another 12% assigned this responsibility primarily towards the employment organization. This indicates that many TAWs do not seem to recognize the potential advantage that a TAW construction can have for them: the fact that they have two employment organizations they can ask for OfCD and that can possibly collaborate in facilitating their OfCD. Interestingly, participants have different arguments regarding the client organization’s responsibility for offering OfCD for compared to that of the employment agency. Whereas the employment agency’s responsibility for offering OfCD was primarily related to the official contract and written agreements, the responsibility of the client organization was primarily related to aspects of the psychological contract or the unwritten agreements, in return for TAWs’ contribution to, and visibility within, this client organization.

The results of our qualitative analyses regarding the division of responsibilities of OfCD indeed align with the results of the quantitative analyses regarding the relationship between TAWs’ expectations and fulfilled expectations of OfCD and their AC towards both organizations. They have higher expectations from the client organization regarding OfCD and also assign the client organization as primarily responsible for offering them opportunities. However, the qualitative analyses also showed a nuance on this result, in that only a minority of the participants also considered it their own responsibility (4%), or partly their own responsibility (6%), to seek OfCD. Still, these results contradict with previous research by Veld and colleagues (2015) who concluded that employees perceived offering OfCD to be a shared responsibility of both organizations and employees. The current study indicated that TAWs, being involved in a triangular relationship, do not see this shared responsibility as such, as only a small minority appointed to their own role in looking for these opportunities.

Overall conclusions

All results of this study together revealed that TAWs tend to be primarily focused on their client organization; it seems as if TAWs consider the psychological contract with the client organization to be more important. This is the organization that most of them regard as responsible for offering them OfCD, and also from which most of them expect to get OfCD. This is the organization for which TAWs have the clearest expectations with regard to OfCD, and for which the effects (explained variances) of expectations and fulfilled expectations of OfCD on TAWs’ AC are the largest. TAWs possibly relate CD directly with their work at the client organization, so it results in the assumption they will get OfCD from their client organization. This interpretation might also explain this study’s result that primarily parallel effects were found between perceived OfCD and AC towards the corresponding organization. And it may also explain two other results. Firstly, the result that only positive parallel relationships were found between expected and fulfilled OfCD and AC towards the client organization, whereas a negative parallel relationship was also found between fulfilled OfCD and AC towards the employment agency. And secondly, the fact that slightly stronger cross-relationships were found between TAWs’ expectations and fulfilled expectations of OfCD by the employment agency and their AC towards the client organization, as compared to the relationship the other way around.

TAWs possibly feel more familiar with the client organization compared to the employment agency, due to the fact that they are physically present at the client organization most of their time. This may result in TAWs feeling ambiguous towards the employment agency, and therefore having a somewhat hesitant attitude. As such, it seems plausible that exceeded expectations of OfCD offered by the client organization contribute positively to TAWs’ AC towards the client organization (all that is familiar is taken for granted, and this needs to be exceeded to stand out, cf. Obuya and Rugimbana 2014), whereas TAWs’ AC towards the employment agency is easily negatively influenced by aspects such as OfCD (all that disappoints them impairs their AC towards the employment agency). Still, the result that expectations of offered OfCD by the employment agency contributed positively to TAWs AC towards the client organization shows that the role of the employment agency should not be underestimated. This result stresses the potential of the employment agency in offering OfCD, as it does seem to pay off with regard to AC towards the client organization.

Based on our data, it can be argued that TAWs do not know what they can expect from the employment agency when it comes to OfCD. Our qualitative data indicated that TAWs perceive this agency primarily as an institution that takes care of the official contract and the associated written agreements without further warmth, emotion and commitment, and not taking into account facilities for competence development. One explanation for this result might be that, at the moment of data collection, the employment agency in question had just started offering broader OfCD to talented TAWs. As such, it is plausible that many TAWs were not yet familiar with the idea of getting OfCD from their employment agency. However, another explanation for this phenomenon that is in our view even more plausible is that TAWs might not be aware of their position in a triangular relationship and what this situation can bring them. With a primary focus on the client organization, TAWs may see the employment agency as an intermediate agent (cf. Morf, Arnhold, and Staffelbach 2014) that sometimes may even interfere with their relationship with the client organization, rather than an employer with which they can discuss OfCD.

This latter explanation fits in with a ‘traditional’ perception of temporary work, with temporary workers being hired to cover short-term absence of permanent staff (cf. Von Hippel et al. 1997). However, times have changed and we believe that this traditional picture does not cover the current real-world situation. As described by Håkansson and Isidorsson (2015), the line between TAWs and permanent employees is becoming blurred these days as organizations increasingly hire TAWs for the long/longer term and for jobs that are equal to those done by permanent staff. This also often applied for the participants in our study; the majority of them were employed at a client organization for one year or more, making them a quasi-permanent part of the client organization. As a consequence, we recognize a trend, with client organizations becoming increasingly aware of the fact that they have to offer equal opportunities to their TAWs and their permanent employees (cf. Mittlacher 2008); employment agencies are also increasingly aware of their role with regard to OfCD. At the same time, TAWs could take their own responsibility in initiating a discussion about the benefits of a TAW construction for them, and the OfCD they get from both of their employment organizations. Strictly speaking, as TAWs need to maintain their market value, they have a responsibility to continuously develop themselves.

Scientific and practical implications

Given the importance of TAWs and TAW constructions for our economies, we believe that we, as HRD researchers and HRD practitioners, need to take the concept of TAW constructions very seriously. Only if both the TAW and the two employment organizations involved feel that they can win something from this situation would a TAW construction indeed be beneficial for all parties, ultimately bringing positive results. Apart from the obvious financial profits for the employment organization and the client organization, this study confirmed that TAWs consider getting OfCD to be positive and profitable for themselves.

With regard to scientific HRD research focused on the triangular relationship of TAW constructions, this study contributed to theory building of the mutual relationships among TAWs, their client organization and the employment agency. Based on TAWs’ AC and the (fulfilled) expectations of OfCD they experienced, this study showed that the triangular relationship is not in balance yet. Instead of functioning as an actual triangle with an equal role of two employers, TAWs perceive their employment construction as primary having a relationship with their client organization. The employment agency, instead, is perceived as being a disturbing factor on this relationship rather than that it is one of their employers with all its related responsibilities and opportunities (Figure 2).

Figure 2. Mutual relationships of a TAW construction in terms of affective commitment (AC) and TAWs’ (fulfilled) expectations of opportunities for competence development (OfCD), as resulted from this study.

Note: The relationship between the employment agency and the client organization (i.e. the business relationship, dotted line) was not part of this study.

As such, this study showed that there is much potential in TAW constructions that is not used yet. Future research may further build on these observations by examining how the triangular relationship can be improved so that all parties involved strengthen each other and gain maximum profit from this promising employment relationship. With regard to HRD, this could imply that research focuses on exploring which instruments help organizations and TAWs to communicate with one another about each other’s expectations and strategies with regard to OfCD. That is, instruments may help to monitor and manage TAWs’ competence development so that it (1) contributes to TAWs’ employability, that it (2) aligns with the business strategies of a client organization and that it (3) contributes to the employment agency’s goals as well.

Besides, future HRD research focused on TAWs may take into account the fact that TAWs have different expectations of OfCD of both of their employment parties as well as that the extent to which these expectations are fulfilled have different effects on the AC towards the employment agency compared to their AC towards the client organization.

With regard to organizational practice, we believe that HRD practitioners can play a crucial role in creating a link between TAWs and their two employment organizations, in turn facilitating a beneficial triangular employment relationship. Specifically, HRD practitioners – both HRD practitioners that work for client organizations and HRD practitioners that work for employment agencies – may actively promote the use of OfCD, rather than using a passive approach that depends on the employee him/herself asking for opportunities. By actively communicating what OfCD a TAW has at both organizations to both TAWs themselves and to their direct managers, TAWs may become more aware of these OfCD, which may result in them making more use of these opportunities. We therefore believe that HRD practitioners can stimulate a TAW to actually make use of these opportunities, eventually contributing to their AC.

As mentioned earlier in this section, we also believe that HRD practitioners can contribute to TAWs’ employability by focusing more on employee benefits rather than taking organizational performances as a starting point in their policies. Although this may involve a cultural change in client organizations which cannot be achieved right away, we believe that HRD practitioners can contribute to TAWs’ employability immediately through tangible measures. Examples include offering them help with their career orientation (either inside or outside their current client organization), organizing workshops to work on skills (such as networking skills) that are extra valuable when working in TAW constructions, and organizing or facilitating meetings between TAWs (either TAWs working at the same client organization, or TAWs having similar functions and working in different client organizations) to share their experiences and dilemmas or to share their expertise and to expand their network.

Limitations and suggestions for future research

The present study has some limitations that need to be addressed. Firstly, this study was built on cross-sectional data. Despite the fact that we were able to draw conclusions on the relationships between offered OfCD and TAWs’ level of AC, we were not able to draw conclusions on causality among these. Although we can see the difficulty in gathering longitudinal data among TAWs, as they may often shift between organizations, we believe that longitudinal data may provide additional information to confirm, or nuance, the results of the present study. Secondly, we need to be cautious about generalizing the results as the TAWs in this study all worked through the same employment agency, and also because of the low response rate in our study (6.7%). Although our sample represented the TAW population of the employment agency in question very well, it was not representative of the total Dutch TAW population. Specifically, our sample included more highly educated TAWs than the ratio of highly educated TAWs in the total population. An explanation for the low response rate can be found in the results of our study. TAWs appeared to be more affectively committed towards their client organization than to their employment agency. Perhaps they did not feel any urgency to fill out this survey as it was related to their employment agency, and because they did not feel really committed to their employment agency. As the low response rate was in line with the response rates of employee satisfaction surveys disseminated annually by the employment agency in question, and also with other studies focused on TAWs, such as Morf, Arnhold, and Staffelbach (2014), we were confident to present the results, albeit exploratory.

Thirdly, in the present study, expectations and fulfilled expectations of OfCD were measured by means of one question per variable, with ordinal answer categories. As a result, more traditional data analysis methods were used. Future studies may use more questions per construct and questions with at least five answer categories, to allow for more advanced data analysis techniques. Nevertheless, given the exploratory character of this study, it presented relevant new insights into the relationship between TAWs’ expectations and fulfilled expectations of OfCD, and their AC towards both employment organizations.

In addition to these limitations, we have additional suggestions for HRD researchers whose work focuses on TAWs. First of all, it would be interesting to nuance the conclusions of this study by differentiating among TAWs. For the present study, we deliberately chose to conduct a large-scale study and to consider all TAWs as one group, in order to reveal a general picture of the situation. By examining the AC of TAWs towards both of their employment organizations simultaneously, and by examining the relationship between perceived OfCD and AC, this study responded to the gap in TAW-oriented scientific research about organizational commitment and competence development. Nevertheless, we believe that the archetypal TAW does not exist, and given the rather poorly explained variances found in this study, we believe that a differentiated picture can provide interesting additional information. Differences might exist among TAWs relating to the kinds of work they do (jobs/roles), the degree to which they are financially (in)dependent, the size of their contract with the employment agency and the background they have with the client organizations. As such, future research may investigate the extent to which different ‘profiles’ can be distinguished among TAWs, facilitating better differentiation between the various TAW constructions. Differentiating among TAWs, and among TAW constructions, would make it possible to create a maximally beneficial situation for the TAW himself/herself, and for the two employment organizations.

Furthermore, future research may also include the perspectives of the employment agency and the client organization, in addition to the perspectives of TAWs themselves. In this study, we deliberately chose to focus on the perspective of TAWs, as this perspective has been underexposed in TAW research to date (e.g. Soltani and Wilkinson 2010). In order to create a balanced TAW construction, however, information must be gained from all the parties involved, and these parties must be brought together to discuss their wishes, demands, conditions and other practical issues regarding OfCD.

Organizations worldwide are increasingly expanding their flexible workforces, in which TAWs often dominate. Let us take care of them very well, both in research and in practice. Offering OfCD has the potential to create a beneficial and profitable situation for all three parties involved in a TAW construction, including the TAWs themselves, and the role of employment agencies should not be underestimated in this respect.

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

1. Interactions between TAWs’ expectations and fulfilled expectations were also measured. However, none of them showed significant effects. We therefore decided not to report them. The results of these interactions can be obtained from the first author.

2. See note 1.

3. See note 1.

4. See note 1.