Abstract
The Work Profiler is an instrument that indicates a job seeker’s probability of work resumption within a year. In addition, it offers a quick diagnosis of the job seeker’s obstacles for return to work. The Work Profiler was developed through a three-stage research process (a literature review, a cross-sectional study and a longitudinal study) in order to identify the best predictive factors for work resumption. While a large part of the research was done regionally in the Province of North Holland, the Dutch Institute for Employee Benefit Schemes (UWV) has decided to use it nationwide. At present this profiling instrument is already being used in almost a third of all unemployment offices in the Netherlands.
Introduction
Due to government-imposed cost cuts, clients on unemployment benefits in the Netherlands will increasingly receive computerised services instead of face-to-face services. The instrument, called the ‘Work Profiler’, enables the selection of job seekers towards face-to-face or computerised services by estimating the clients’ chances of finding work within a year. In addition, it offers a quick diagnosis of the most important obstacles for their return to work, thereby determining the type of services needed to enhance a client’s probabilities of resuming work.
This paper aims to provide a clear picture of the Work Profiler. At this moment, the instrument has already been introduced at 11 out of the 35 unemployment offices in the Netherlands. While all unemployment offices operate locally, as a whole they form the Public Employment Services, which is part of the Dutch Institute for Employee Benefit Schemes (from hereon referred to by its Dutch abbreviated name ‘UWV’). In 2013, the 35 unemployment offices aided over 600,000 persons who recently lost their jobs and applied for an unemployment benefit.
Exploring the Work Profiler
The Work Profiler is a selection and diagnosis instrument which aids the UWV to provide tailored service provision to clients on unemployment benefits. To this aim, a client fills in a short questionnaire containing 20 questions within the first three months of unemployment. The 20 questions relate to 11 hard and soft factors that are predictive of return to work. The Work Profiler provides two outcomes based on the answers given by the client. The first outcome shows the client’s chance of resuming work within one year. The second outcome provides a quick diagnosis by illustrating which of the 11 predictive factors for work resumption need to be positively influenced in order to increase the client’s chances of returning to work.
The two outcomes of the Work Profiler will be used by UWV for two purposes:
Selection: The client’s chance of work resumption within a year will be employed to determine whether the jobseeker will be offered computerised or face-to-face services. Due to government-imposed cost cuts, UWV is no longer able to offer face-to-face services to every person on unemployment benefits. The Work Profiler provides an objective method to select those eligible for face-to-face services. The remainder of clients will primarily make use of computerised service provision, which is available to all. Quick diagnosis: Disregarding whether a client will use computerised or face-to-face services, it remains important to offer services which will increase a client’s chances of reemployment. UWV utilises the second outcome of the Work Profiler, a quick diagnosis based upon the 11 predictive factors for work resumption, for this purpose. The scores on the individual factors indicate the client’s strong points and which may need improvement in order to enhance a client’s chances for finding work. By offering services which aim is to act on those factors that need enhancement, UWV is able to provide tailored services. Such services can be offered both in face-to-face contact as well as in a computerised environment.
Operating context of the Work Profiler
In 2006, UWV set out to create an instrument to aid professionals at unemployment offices in their contact with job seekers by identifying the predictors for work resumption. At the time the unemployed only made use of face-to-face services. However, by the time the research was completed and the Work Profiler was created, much had changed. Cost cuts imposed by the Dutch government made UWV decide to switch from face-to-face to computerised services as the main channel of contact. Nowadays, all job seekers on unemployment benefits (which are able) make use of computerised services and only a small portion additionally gets face-to-face services. The Work Profiler was not developed originally for a digital environment, but proved adaptable and now operates within a computerised environment.
Because UWV switched to digital services only a few years ago, much development and implementation work is still being carried out. Nevertheless, their introduction has been successful and today the great majority of the unemployed use exclusively computer-based services. Every unemployed person has access to his/her own individual digital environment through an online portal. This personalised environment is called Work Folder or Werkmap in Dutch (Figure 1). The Work Folder can be considered a personal file managed by the client, where he will find online modules that aid in the search for work, such as webinars and e-learning programs in how to write an application letter, make a curriculum, prepare for a job interview, or look for vacancies that correspond with their profile. It is also his primary channel of communication with the work professionals of the unemployment office. A client can, for example, upload documents (curriculum or letters) and choose to share them with their work professional in order to get feedback.
The online personal environment of the Work Folder, where job seekers will find the questionnaire underpinning the Work Profiler (translated screenshot).
The Work Profiler is one of various tasks inside the digital environment of the Work Folder which the client is asked to complete at the start of unemployment. Based upon the outcome, the client may be given additional face-to-face services.
The contents of the Work Profiler
Factors and corresponding questions that constitute the Work Profiler.
Outcome 1: Chances of work resumption within one year
The Work Profiler indicates the chances of work resumption through a percentage, which, as shown below, is 0% being the lowest and 100% being the highest probability for return to work (Figure 2). UWV can then use the indicated probability as a selection criterion to determine which clients will receive face-to-face services. Job seekers with poor prospects in the labour market who have been predicted not to resume work within a year may seem a logical choice for face-to-face services. However, that is not the case, because at present there are much more job seekers with poor employment prospects than there are face-to-face services available. The Work Profiler can thus be used to select a small number of clients with poor work prospects to fit the capacity of the face-to-face services by those who fall within a certain margin of possibility of returning to work. For instance, all who fall below, above or between some percentage.
The two outcomes of the Work Profiler as seen by the work professional (translated and simplified screenshot).
One of the main advantages of the Work Profiler is that even if the available budget for helping the unemployed find work may change with time, the instrument remains adaptable. For example, if in the future the budget should increase, a larger number of unemployed can be offered face-to-face services. Their selection can be done by looking at their ‘changes of work resumption’ according to the Work Profiler and broadening (or narrowing in case of cost cuts) the section of percentages that lead to face-to-face services.
Outcome 2: Insights into which factors to influence
UWV is able to make a quick diagnosis of the client’s situation at the start of his unemployment through the individual scores on the 11 factors. This then forms the basis for UWV to offer a (computerised) service that takes into account the client’s personal obstacles (Figure 2).
Since the Work Profiler was developed by UWV to be used for selection and quick diagnosis, its short questionnaire only focuses on the most important predictors for work resumption. Hence, the Work Profiler does not offer a total personal profile with in-depth insights into the competencies and capacities for return to work. To this aim, UWV has more extensive tests done by its specialised ‘Competencies Centres’.
The development of the Work Profiler
In order to create the Work Profiler, we first needed to identify which factors are predictive for return to work at the start of unemployment. An elaborate study called ‘Predictors of Work Resumption’ which aimed to identify these factors preceded its creation. The research project was a collaboration between the UWV Centre for Knowledge (Kenniscentrum UWV) and the School of Medical Sciences of the University Medical Centre Groningen (UMCG) and was carried out from 2006 to 2011. A detailed description of the research project can be found in Brouwer et al. (2011).
The central question of the study was: what are the predictive characteristics of the persons and/or their personal situation for resuming work within 12 months after becoming unemployed? The aim was to identify the characteristics which are predictive at the start of unemployment, since this would allow UWV at an early stage to recognise persons who run a high risk of long-term unemployment. Early recognition also allows UWV to offer services or interventions which aim to enhance the probabilities of a client’s work resumption during the early phase of unemployment.
Hard and soft factors
When looking at possible predictive factors for work resumption, a distinction can be made between ‘hard’ and ‘soft’ factors. In general, the influence of hard factors such as age, education or having a disability is well known. Such data are usually easily available for research since those are typical client characteristics that tend to be included in the administration of an unemployment benefits organisation. During the early stages of the research project, it became clear that very little light has been shed upon the soft factors that entail a person’s psycho-social situation in relationship to work resumption, such as job search intention, job search behaviour, motivation and perceived health. We expected that these factors would indeed influence a client’s chances in the labour market (e.g. Kanfer et al., 2001; Saks, 2005; Thomsen, 2009), but there was only limited literature to support this notion (Gelderblom et al., 2007). One of the additional aims of the research project was to look at the predictive value of soft factors and to incorporate them with the predictive hard factors into a single instrument.
Three stages of research
The research project ‘Predictors of Work Resumption’ was carried out in three subsequent stages: a literature study, followed by a cross-sectional study, and finally a longitudinal study. By this process, we were able to reduce the large number of possible relevant factors for work resumption mentioned in the literature to a much smaller and predictive set. Figure 3 gives a schematic representation of the funnel process during these three stages. This section offers the most essential information. In Brouwer et al. (2011), an elaborate description of the methodology and considerations can be found.
The research process in three stages.
The initial extensive literature study produced many hard and soft factors that may influence work resumption. From the literature, three theoretical models that encompassed these factors were selected, which are the Wanberg model (Wanberg et al., 2002), the Theory of Planned Behaviour (Ajzen, 1991) and the Expectancy Theory of Motivation (Vroom, 1964). A total of 500 items were found relevant, which were used during the next stage.
The 500 items were processed into a questionnaire for the purpose of the cross-sectional study, which compared the answers of the long-termed unemployed with those of whom resumed work quickly. By this process, the number of items was reduced from 500 to 155. The nature of cross-sectional research itself did not allow establishing which factors are predictive; it could only demonstrate the differences between the two populations.
The true predictive value of factors could only be established through a longitudinal study, which was done during the last stage of the research process. In this phase, a large group of recently unemployed people was asked to fill in a questionnaire containing the now reduced list of 155 items. A year later, a follow-up was done of the exact same people in the administration of the unemployment office to see whether they did or did not resume work. The longitudinal study was carried out in the Province of North Holland from April 2008 to March 2009 among all those who became unemployed. The total research population consisted of 3618 respondents of which 58% found work within a year and 42% still had an unemployment benefit after one year. Job seekers whose unemployment benefits ended before 12 months for causes other than work resumption, such as reaching the maximum duration of their benefit, were excluded from the analysis. The longitudinal study allowed for a further reduction of items from 155 to 20.
Predictors of work resumption
Out of more than 70 factors, only 10 resulted as predictive for work resumption within 12 months, as indicated during the last stage of the research by multivariate logistic regression analysis. In other words, to get a good grasp upon the chances for finding work one only has to look at these 10 factors. The factor ‘physical work ability’ did not prove predictive in the multivariate logistic regression analysis, but was added upon request from the work professionals. They felt that it was a necessary counterbalance of the predictive factor ‘mental work ability’ in order to get a complete picture of the job seeker’s health perception (Havinga and Hijlkema, 2012).
Of the final 11 factors, six enhance and five reduce a person’s probabilities for work resumption. The chances of finding work will increase when the job seeker has positive views on returning to work, has already had contact with employers, has a positive job search intention, and has an overall positively perceived general, physical and mental work ability. Factors that hinder work resumption are: an older age, having been employed for many years in the same job, having problems understanding Dutch, feeling too ill to work, and having a tendency to ascribe the unemployment to external coincidental circumstances (i.e. a high external variable attribution). Table 1 offers an overview of the 11 factors and their corresponding 20 items.
The results of the research process were used to create the Work Profiler. Its questions have been formulated in such a manner that they are easy to understand, even by people with relatively poor Dutch language skills. The two outcomes of the Work Profiler, which UWV uses for selection and for a quick diagnosis, as mentioned above, are calculated through formulas established during the last phase of the research process.
Outcome 1: The chances of resuming work within one year (selection)
This outcome is based on a formula that uses the results of the multivariate logistic regression analysis to calculate the chances of resuming work within 12 months on an individual basis. This chance is indicated by a percentage (0% = no work resumption; 100% = work resumption). The probability of work resumption can be used as a means to select which clients are eligible for face-to-face services.
Outcome 2: Factors to be influenced (quick diagnosis)
The Work Profiler offers a quick diagnosis of a client’s situation by looking at which factors may be influenced in order to enhance the probabilities of work resumption. This is done by comparing a client’s individual score to the average scores of the reference group, which are all jobseekers from the research population who returned to work within a year. For each of the 11 factors such a comparison is made and the Work Profiler ends up with a personalised indication per factor which is expressed as above average, average or below average. An above average indication is beneficial for some factors (e.g. views on return to work and job search intention), while disadvantageous for others, where a below average is desired (e.g. problems understanding Dutch and feeling too ill to work). Since the Work Profiler is used at an early stage of unemployment (i.e. the first three months) the outcomes of the quick diagnosis can be used especially to positively influence those factors that need further attention to enhance a person’s possibilities for return to work.
Currently the Work Profiler predicts with a certainty of 70%. This means that during the first few months of unemployment the Work Profiler is able to determine correctly for 7 out of each 10 jobseekers who will resume work within a year.
Initial experiences from job seekers
Between May and December 2013, the Work Profiler has been implemented stepwise in 11 Dutch unemployment offices out of the total 35. Clients from these offices were approached in April 2014 to take part in an assessment study. They were asked whether the aim of the instrument was clear and what they thought of its user friendliness. In total, 3300 clients were invited of whom 1290 (39%) participated in an online assessment.
On average, clients gave the Work Profiler 7 out of 10. Most of them were positive and stated that it was easy to operate the instrument online, took them little time and was easy to understand. Periodical monitoring indeed demonstrates that clients on average need only a few minutes to answer the questions. Often job seekers indicated that the instrument helped them become aware of their situation and allowed them to better organise their consequent actions in looking for work.
The least positive conclusion was that approximately a quarter of the respondents did not know what to expect after filling in the Work Profiler questionnaire. On the one hand, this has led UWV to decide to make changes in its communication with clients immediately prior and after using the Work Profiler. On the other hand, it underscores the current ongoing development of connecting the Work Profiler’s outcomes with computerised services.
Ongoing development of the Work Profiler
Continuous development on the Work Profiler is needed to maintain the instrument up-to-date and reliable. For this reason, while simultaneously implementing the Work Profiler (version 1.0) on a national scale, a follow-up study has also started that will result in version 2.0. The former version was based upon the influx cohort of the Province of North Holland, whereas the latter study will take place throughout the nation. Besides verifying which factors are (still) predictive for work resumption within a year, the follow-up study additionally aims to enable the Work Profiler to estimate the duration of the unemployment.
In broad terms, the follow-up study can be considered as a partial repetition of the third stage (i.e. the longitudinal study) of the earlier investigation process. The list of 20 questions from the Work Profiler 1.0 has extended to 55 items. Starting from March 2014, during a yearlong period a digital questionnaire containing these items will be offered online to all clients at 11 unemployment offices scattered throughout the Netherlands. Subsequently, it will be possible to establish, from UWV administration, whether these clients have indeed resumed work or not. The resulting predictive factors and corresponding items will form the core of the Work Profiler 2.0 that will replace the older version in 2017.
Footnotes
Funding
This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.