Does receipt of unemployment benefits change recruiter perceptions of candidates’ personality,work relevant skills and employability?

Abstract

BACKGROUND:

Growing experimental evidence shows that unemployment benefit recipients are generally perceived negatively in terms of their personality and employability by the general public. Welfare stigma tied to unemployment or receipt of income support may disproportionately negatively impact individuals who have been out of work due to disability, or chronic health conditions.

OBJECTIVE:

The current study examined whether welfare stigma and/or unemployment stigma, translate to perceptions and hiring decisions made by individuals working in recruitment, potentially creating barriers to re-employment for those without work and relying on unemployment benefits.

METHODS:

We used a vignette-based experiment (N = 213) where participants working in recruitment rated personality and employment capabilities of characters who were described as employed, unemployed or unemployed and receiving benefits.

RESULTS:

Characters who were employed were generally rated more positively on employability and work-relevant skills, compared to the unemployed and unemployed benefit recipients, but these differences did not translate into a binary hiring decision (would you hire this person for the job). There were few differences in ratings of personality characteristics between the employed, the unemployed and those who were receiving unemployment benefits.

CONCLUSIONS:

These results add to knowledge about the determinants of welfare stigma showing that potential bias towards the unemployed and benefit recipients held by recruiters differs from that held by the general public.

Keywords

Employability unemployment personality big five vignette

1 Introduction

A major ongoing challenge facing governments globally is ensuring that citizens in prime working-age can participate in social and economic life [1], a task that will be increasingly prominent in the economic recovery from the COVID-19 crisis. Building better pathways to participation, particularly for individuals receiving unemployment benefits, including those with disability, has been the focus of welfare-to-work programs and policies across the Organisation for Economic Co-operation and Development (OECD) countries since the 80s, including Australia [2, 3]. Research literature from Australia and the US identifies common barriers that make moving from unemployment benefits to employment difficult, including low basic skill level, lack of recent work experience, lack of ‘work readiness’ or ‘soft skills’, and experiences of employer discrimination [4, 5]. Particularly the stigma and potential discrimination associated with unemployment benefit are known to have negative impacts on health, employability and equality (for meta-analyses, see [6]). However, existing literature often conflates the stigma due to unemployment and stigma due to welfare receipt, that are likely to represent separate –but related –constructs [7]. The aim of the current study was specifically to separate and examine the impact of the two different types and co-occurring stigmas on perceptions of individuals who make recruitment decisions.

The stigma attached to the receipt of unemploy-ment benefits is also a potential barrier to labour-market success [8]. General population surveys show that unemployed benefit recipients are seen as “lazy and dependent” in Australia and the UK [9 –12], or hostile, ill-intentioned and incompetent in the USA [13] in comparison to individuals who are not receiving unemployment benefits. Other research methods demonstrate an unconscious effect of stereotypes towards unemployed benefit recipients that can be differentiated from explicit attitudes that are likely to play a role in judgments about the support unemployed people should receive from the welfare system [7]. However, these studies did not determine whether the negative perceptions were attributable to the state of unemployment, receiving benefits, or both. These negative stereotypes, both implicit and explicit, can be considered to indicate ‘welfare stigma’, and discourage take-up of benefits, discrimination and adverse effects of recipients’ perceptions of themselves [14, 15]. In a recruitment context the implicit biases attached to welfare stigma may also lead to discrimination in hiring decisions, creating further barriers to (re)employment.

Previous studies on bias in employment decisions typically involve generating resumes for fictitious candidates that vary with respect to the appearance, race or ethnicity of the applicants based on name or photos [16, 17]. These resumes are submitted to advertised positions and differences in call-backs (i.e., invitations to job interviews) or the final hiring decisions [18] provide a measure of discrimination. These types of field experiments show that employers base their hiring decisions not only on the job selection criteria or curriculum vitaes (CVs) but also on stereotypes —that is, category-based traits or attributes that are often applied to a group of people as a result of accepted beliefs about the members of the group that can be implicit or explicit [19, 20]. Group stereotypes lead to expectations about how members of the group, such as unemployment benefit recipients, should and do behave. Although they can be functional, automatic, unintentional, and accurate in the aggregate [21, 22], stereotypes can result in bias: an inaccurate evaluation reflecting a generalization rather than an individual’s true qualities. While we have found no research examining job success rates based on unemployment benefit receipt, studies that manipulate current employment status on resumes show a lower rate of call-backs and job search success for those who are unemployed, compared to employed candidates [23, 24]; this being the case particularly when unemployment was current and long term (as opposed historical or short term).

While few studies focus on perceptions of other’s employability, perceptions of one’s own employability have been recently studied in university students, jobseekers [25], the unemployed [26] and currently employed workers [27, 28]. These studies consistently show that higher levels of perceived employability are linked to personal and job-related wellbeing as well as career success. Examining others’ perceptions of employability, particularly those who make recruitment decisions, may be more relevant to understand factors impacting on actual employment outcomes. However, the existing studies examining other’s perceptions of employability have focused exclusively on job specific skills [29, 30], and not on perceptions or stereotypes based on a group membership, such as being unemployed or receiving unemployment benefits.

Stereotyping of certain groups is more likely in the absence of other differentiating information. The more information a decision maker has about a candidates’ credentials, skills, relevant experience, and the like, the less they rely on group status, (benefit or employment status) as the basis for their decision [31]. Statistical or stereotype discrimination concerns estimates about characteristics that are unobserved on CVs but correlated with group affiliation [17]. Thus, groups can be treated differently for many reasons ranging from arbitrary preferences for or against a group to differentiated beliefs about various groups’ (including benefit recipients) productivity-related characteristics [32]. Very small differences in stereotypes, or arbitrary preferences, held by employers can lead to large differences in call-backs [33].

Welfare and benefit recipients are generally viewed negatively, and the strength of these negative stereotypes vary only somewhat across nations [34] and across the reasons for government assistance (e.g., unemployment, disability, lone parent status; [35, 36]. Unemployed benefit recipients are typically seen as the group least deserving of government support, being more likely to be seen as responsible for their own plight, ungrateful for support, and not in genuine need. They are also perceived as low in reciprocity [37, 38] which reflects that they are seen as taking more than they have given –or will give –back to society [39]. These types of appraisals, if made in a hiring context, would disadvantage those receiving welfare benefits compared to otherwise similar candidates not receiving benefits.

A number of experimental paradigms have been used to investigate perceptions of “welfare recipients” and the “unemployed”. The stereotype content model (SCM; [14]), for example, represents the stereotypes of social groups on two dimensions: warmth, relating to being friendly and well intentioned (rather than ill intentioned); and competence, relating to one’s capacity to pursue intentions [14]. Using this model, the “unemployed” have been evaluated as low in warmth and competence across a variety of welfare regime types [14]. The structure of stereotypes has also been studied using the Big Five personality dimensions [40 –42] Openness, Conscientiousness, Extraversion, Agreeableness, and Emotional Stability (for background on the Big Five see: [43 –48]. There are parallels between the Big Five and the SCM: warmth relating to the dimension of Agreeableness, and competence relating to Conscientiousness [49 –52] and these constructs have been found to predict employability and career success [53, 54].

Experimental vignette studies have helped to und-erstand the evaluative mechanisms underlying perceptions, showing that the unemployed are perceived more negatively than the employed; specifically, as less employable and as possessing fewer desirable traits [55, 56]. A small body of research using vignette methodology exists on employer’s hiring decisions based on age/receiving age pension [57, 58]. However, we have found no research using this methodology that examines receipt of government benefits. The existing research, therefore, provides no insights into the effect of welfare stigma.

Building on this previous work, our own research has focused on the effects of unemployment by drawing on frameworks of Big Five, SCM and employability in the general population [40 –42]. Our studies consistently show that unemployed individuals receiving government payments are perceived as less employable (poorer ‘quality’ workers and less desirable for employment) and less Conscientious compared to the employed and the unempl-oyed (no reference to benefit status). We found similar but weaker pattern related to Agreeableness, Emotional Stability, and the extent that a person is perceived as ‘uniquely human’ [40, 41]. These negative perceptions are restricted to those currently receiving benefits. We found that vignette characters described as currently employed but with a history of welfare receipt were indistinguishable from those described as employed and with no reference to benefit receipt [41; see also 23]. No existing work has examined the influence of benefit receipt and/or unemployment on perceptions of personality and employability in a context of employment decisions: which is an essential extension of this research paradigm to better understand the practical implications of welfare stigma as a potential barrier to (re-)employment.

1.1 Current study

The overarching purpose of this study was to examine whether people with responsibility for recruitment and hiring decisions hold negative views of those described as unemployed benefit recipients and those described as unemployed (with no reference to benefit status) similar to the general population, and whether these perceptions translate into hiring decisions. As described in the literature above, the formation of perceptions of individuals receiving unemployment benefits has already been studied in general community samples showing that, overall, the unemployed are perceived more negatively than employed individuals, and that unemployed receiving benefits are perceived more negatively than the unemployed when no reference is made to welfare benefit receipt [40 –42]. Specifically, the study objective was to evaluate how recruitment decisions were influenced by the characteristics (e.g., employment status, and welfare status, age, gender, community involvement, education, qualifications and technical skills) of fictitious vignette characters who had applied for a job. Of particular interest in this study were the current employment and welfare status of the applicant. The inclusion of benefit receipt with a range of other characteristics serves to increase the correspondence of the task to an authentic recruitment process and to mask the key experimental manipulation. Overall, we expected to find unfavourable appraisals of benefit recipients in hiring decisions when compared to individuals not receiving welfare benefits.

Based on our previous work [40 –42] on general population samples, we expected that both employment and welfare status will affect employment-related measures including applicant overall quality, ‘employability’ (i.e., whether you would hire this person) and perceptions work-relevant soft skills (e.g., organizational skills) not described in the vignette. Specifically, employed would be rated higher than unemployed (with no reference to benefit status) who will be rated higher than welfare benefit recipients on these dimensions. We also expected that, compared to an employed character, characters receiving government benefits due to their unemployment would be evaluated as less Conscientious (a trait associated with incompetence and laziness [59, 60] and less Agreeable and Emotionally Stable (traits which have been associated with low warmth, and typically co-occur) [49]. Our re-registered hypotheses can be accessed: https://osf.io/uhq5t.

Perceptions of low employability are a barrier to returning to work, and building greater understanding of these factors is essential to inform strategies to break cycles of unemployment and government dependence. In this experimental design, the character’s unemployment status was independent of a range of other work-related characteristics included in the vignette (current benefit recipient status, volunteering, work experience, appearance, work relevant skills).

2 Method

2.1 Sampling and participants

Participants were recruited through an Australian online panel managed by Qualtrics, a market research company which provides both an online survey tool and access to a participant pool from which to recruit. Each participant has signed up (i.e., opted-in) to being invited to participate in research and other online tasks and activities. We sampled from participants who were identified as being employed and involved in making recruitment decisions. The Qualtrics used a quota sampling approach until all experimental conditions were assigned. The final sample were 213 individuals with complete data who met the inclusion criteria: they were 18 years of age or older, currently employed and had been involved in hiring candidates/filling a job vacancy in the previous 12 months (including Employers, HR Managers, and Recruitment specialists). The study participants were all recruited from Australia, aged between 18 and 68 years of age years (M = 41.2; SD = 11.87) and 39.0% were males.

2.2 Main outcomes

2.2.1 Personality

The Ten Item Personality Inventory [61] Ratings of vignette characters on the Big Five were performed using a 10-item adaptation of Saucier’s Mini-Marker scale [62]. This scale contains two items for each trait in the Big Five (Openness, Conscientiousness, Extraversion; Agreeableness; Emotional stability), and each of these items contains a pair of trait adjectives (e.g., “dependable, self disciplined”). The participants were asked to rate the character on each trait using a 5-point scale from “does not describe them at all” to “describes them very well”. When used for ratings of other people, these subscales all exhibit adequate reliability [40 –42].

2.2.2 Employment-related outcomes

All the measures were included as a part of planned (preregistered) exploratory modelling to examine how similar and highly correlated measures hang together, i.e., would the ratings of worker quality translate into hiring decisions. Worker calibre ranging from 0 to 100 was included as an initial measure that would be sensitive to small changes. Then, employability was measured on a Likert scale as it was unclear at what point worker calibre ratings would render a character employable. Finally, the binary variable of the ‘hiring decision’ was included to examine whether assessments of worker calibre and employability would translate into actual hiring decisions. The hiring decision question was presented after all other items, and the process of responding to these items may have helped to consolidate the image of the character in the brief vignette.

Worker calibre , indicating the perceptions of overall worker quality, was measured by a single question: “On a scale from 0–100, where 0 is would not employ even if the only applicant, and 100 is the perfect applicant, what kind of score would you give [character].”

Employability was assessed by a single item. The statement: “I would employ the [character]” was presented, with response options on a 5-point Likert-like scale ranging from ‘strongly agree’ to ‘strongly disagree’, referred to as the employability outcome.

Soft Skills were assessed by a four question scale that was developed for this study on the basis of previous extensive research on soft skills [63, 64]: I expect that [character] has good communication skills; I expect that [character] has good team work skills; I expect that [character] has good leadership skills; I expect that [character] has good organisational skills. Response options for each question ranged on a 5-point Likert-like scale from ‘strongly agree’ to ‘strongly disagree (α= 0.79). These four dimensions have been consistently linked to be indicative of ‘soft skills’ and workplace performance [63, 64].

Hiring decision , indicating the participants’ willingness to hire the character, was assessed by a single question: “Would you hire [character]?” with a binary yes/no response option.

2.3 Procedure

The vignette character rating experiment involved two steps to approximate a recruitment process and to mask experimental aims. The initial phase of the 2-stage design provided participants with a sense a real-world hiring process (i.e., assessment, short-listing, interview), although no information related to the key experimental conditions (employment status and benefit receipt) were provided until the second step. The methodology used does not precisely emulate the 2-stages of ‘real world’ hiring processes (resume assessment, short-listing, interview) as the attributes used to test the hypotheses are only provided at stage 2, even though these are frequently available in resumes. The 2-step process, however, was used to reduce bias by ensuring that everyone is rating 2 characters at Step 2 they consider to be equally desirable.

At step 1, participants were presented brief information about four fictional (randomly assigned) characters, and asked to select the two they would shortlist for interview. All four characters had similar levels of educational attainment (low or high) and other characteristics such as age (25, 30 or 35), gender (female or male), education (four levels within each strata) and work experience (job relevant vs irrelevant) were randomly varied. At step 2, study participants were given more details about each of the short-listed characters including their employment status (main manipulation: employed, unemployed or receiving unemployment benefits). The vignette described the interview scenario, including the circumstance of the character’s arrival, and ten characteristics (varied for each vignette and described in Manipulation section below) to prevent participants focussing on any one piece of information. After reading the vignette for the first character, each study participant rated the character’s work related soft-skills (e.g., communication, team work, leadership, organisational skills), their quality and employability, their personality traits (Openness, Conscientiousness, Extraversion, Agreeableness, and Emotional Stability), and made a decision of whether they would be willing to hire this candidate. This process was then repeated for the second short-listed character who was assigned a different employment/benefit status to the first character, with the other ten characteristics randomly assigned. After reading the second vignette, the participant was asked which of the two characters they would choose to recruit (pre-registered as for exploratory use, and not part of this study). At the end of the survey, participants completed demographic questions about themselves and their experience as a recruiter.

The vignette methodology used in this study builds on previous work with general population samples by the authors [40 –42] varying a large number of background characteristics in each scenario. Vignette paradigms are a common way of examining stereotypes as they help ensure that extraneous characteristics are controlled and enable experimental manipulation of personal characteristics. By averaging across the sample members’ perceptions, an understanding of the characteristics that drive stereotypes in that population (here, recruiters) can be obtained.

2.3.1 Manipulation

The main independent variable the experiment manipulated through the vignettes were three levels of employment and benefit status of the vignette character: (1) employed; (2) unemployed; and (3) unemployed and receiving unemployment benefits. The key conditions were operationalised by a string of text embedded within the vignette (employed: “wanted new opportunities that her/his current employment was unable to provide; unemployed: “hadn’t been able to find work since his/her last contract ended”; unemployment benefits: “hadn’t been able to find work and was trying to get off unemployment benefits”).

Ten other vignette characteristics were manipulated between-subjects. Age was specified as either 25, 30 or 35 years (and randomly selected for each vignette) and the character’s gender was randomly varied, with a common name corresponding to each gender selected without replacement from the available options (male: John, Thomas; female: Mary, Rachel). Participants could be presented with two differently named characters of the same gender and/or of the same age. All names were stereotypically White to keep perceived race constant, and prevent participants’ assumptions about character race from varying across key experimental conditions [65, 66]. Other between subject manipulated characteristics included the candidate’s: (1) appearance (desirable vs. poor); (2) passion for their work (high vs. low); (3) punctuality (high vs. low); (4) relevant work experience (yes/no); (5) volunteering (yes/no); (6) statement about one out of two personality dimensions (Openness to Experience vs Emotional Stability; (7) meeting the selection criteria for the position (all vs. most criteria met); and (8) education (high vs. low). To ensure no correlation with the experimental condition, the features of the second vignette were counterbalanced on each dimension except age and gender, which was balanced between-subjects. Each participant was randomly assigned to an ordered vignette pair that no other participant had completed.

2.3.2 Analysis

Random effects linear and logistic regression models were used to assess whether being employed, unemployed or receiving welfare benefits was associated with how people making hiring decisions perceived the calibre, willingness to hire, employability, soft skills and personality dimensions. With 426 vignette observations nested within 213 employers/participants, the data has a multilevel structure and standard regression analysis was not appropriate. Random effects regression models are a common approach to account for data in which observations are clustered within individuals, groups or time [67]. Vignette data—presuming that the randomization of vignettes to a random sample of respondents was effective—fulfil the main prerequisite for the use of random effects models [68]. In the current study, all outcome measures (personality, worker calibre, employability and soft skills, hiring decision,) were repeated for two different scenarios.

To assess whether employment and welfare status were associated with each of the outcome measures, we present the estimated coefficients from a series of random effects linear regression models for the continuous outcomes (the five personality dimensions, worker calibre, employability and soft skills) and the estimated marginal differences from the random effects logistic regression for binary outcome (hiring decision). Unadjusted and adjusted models were performed with pairwise comparisons between the levels of the employment and benefit status measures for all outcomes. We tested for interactions for the main experimental conditions and co-variates, including volunteering status. The methodology for the study, including the analysis plan, were pre-registered.

3 Results

All 213 participants had been involved in recruitment as part of their professional role in the previous 12 months, with 87.9% (n = 187) employed full-time and the rest 12.1% (n = 26) working part-time. Most (n = 187; 87.9%) participants identified as White, Caucasian or European, 10.7% (n = 23) identified as Asian and the remaining 1.4% identified as African, Indigenous or they preferred not to say. They were relatively highly educated, with 35.7% (n = 76) having completed an undergraduate and 37.6% (n = 80) a postgraduate university degree or diploma. Table 1 presents descriptive statistics and bivariate correlations for the nine outcomes, showing a positive association between all the outcome variables (p < 0.001) but with notably higher correlations among the four employment-related outcomes: worker calibre, employability, soft skills and hiring decision and the personality dimension of Conscientiousness.

Table 1
Descriptive statistics and bivariate correlations for nine outcome variables

M(SD) Range 1 2 3 4 5 6 7 8 9

1 Calibre 72.45(15.17) 0–100 1

2 Employability 4.02(0.84) 1–5 0.744 1

3 Soft skills 2.81(0.69) 0–4 0.590 0.679 1

4 Conscientiousness 3.68(0.79) 1–5 0.424 0.435 0.471 1

5 Extraversion 3.33(0.85) 1–5 0.311 0.306 0.375 0.515 1

6 Agreeableness 3.63(0.67) 1–5 0.271 0.252 0.259 0.495 0.301 1

7 Openness 3.43(0.77) 1–5 0.266 0.318 0.339 0.304 0.306 0.257 1

8 Emot. stability 3.41(0.69) 1–5 0.224 0.244 0.233 0.409 0.318 0.392 0.203 1

9 Hiring decision^* Yes: 82.16% 0, 1 0.596 0.745 0.529 0.551 0.336 0.312 0.238 0.309 1

		M(SD)	Range	1	2	3	4	5	6	7	8	9
1	Calibre	72.45(15.17)	0–100	1
2	Employability	4.02(0.84)	1–5	0.744	1
3	Soft skills	2.81(0.69)	0–4	0.590	0.679	1
4	Conscientiousness	3.68(0.79)	1–5	0.424	0.435	0.471	1
5	Extraversion	3.33(0.85)	1–5	0.311	0.306	0.375	0.515	1
6	Agreeableness	3.63(0.67)	1–5	0.271	0.252	0.259	0.495	0.301	1
7	Openness	3.43(0.77)	1–5	0.266	0.318	0.339	0.304	0.306	0.257	1
8	Emot. stability	3.41(0.69)	1–5	0.224	0.244	0.233	0.409	0.318	0.392	0.203	1
9	Hiring decision^*	Yes: 82.16%	0, 1	0.596	0.745	0.529	0.551	0.336	0.312	0.238	0.309	1

^*Polychoric correlation was used to estimate the rho between continuous and binary variable (hiring decision), correlations between continuous outcomes are estimated with Pearson’s correlation.

We then directly compared the three levels of the main experimental variable: (1) employed, (2) unemployed but no reference to benefit status and (3) unemployment benefit conditions on each of the employment- and personality-related outcomes. Unadjusted and adjusted random effects models relating to the three key experimental conditions yielded broadly consistent patterns of results. Consistent with the pre-registration, the unadjusted models are reported here (adjusted models shown in Table S1).

Considering the employment-related outcomes, the comparisons between the three experimental groups in Table 2 show that employed characters were rated higher than both unemployed and benefit-receiving characters on worker ‘calibre’ (z > 2), empl-oyability (z > 3) and soft skills (z > 2), and employed characters were rated higher than only benefit rec-ipients on soft skills (z > 2). While there was a gradient across the three conditions on the esti-mates for these outcome measures (with empl-oyed > unemployed > benefit recipients), there were no statistically significant differences between the unemployed and benefit recipient characters. While the same pattern of results was evident for the binary hiring-decision outcome (with 87% of employed characters hired compared to 80% of unemployed and 79% of benefit recipient characters), these differences were not statistically significant (zs < 2). However, we note that after controlling for the other experimental variables, the difference between the employed and unemployed conditions and between the employed and benefit receipt conditions were both statistically significant. Again, there was no difference between the unemployed characters and those receiving benefits.

Table 2

Unadjusted sequential comparisons for three levels of main condition

	Employed (ref) vs. Benefit			Employed (ref) vs. Unemployed			Unemployed (ref) vs. Benefit
	b (se)	[95% CI]	sig	b (se)	[95% CI]	sig	b (se)	[95% CI]	sig
Unadjusted
Calibre	–5.43 (1.45)	[–8.27, –2.59]	^***	–3.10 (1.21)	[–5.88, –0.32]	^*	–2.33 (1.44)	[–5.14, 0.48]
Employability	–0.26 (0.09)	[–0.44. –0.09]	^**	–0.20 (0.09)	[–0.37, –0.09]	^*	–0.07 (0.09)	[–0.24, 0.11]
Soft skills	–0.17 (0.06)	[–0.30, –0.04]	^**	–0.13 (0.06)	[–0.25, 0.04]	^*	–0.04 (0.07)	[–0.17, 0.08]
Conscientiousness	–0.11 (0.08)	[–0.32, 0.06]		–0.13(0.09)	[–0.31, 0.06]		–0.01 (0.09)	[–0.18, 0.19]
Extraversion	0.15 (0.10)	[–0.05, 0.35]		0.18 (0.10)	[–0.02, 0.38]		0.04 (0.10)	[–0.16, 0.24]
Agreeableness	0.14 (0.07)	[0.01, 0.27]	^*	–0.04 (0.07)	[–0.17, 0.09]		0.18 (0.08)	[0.05, 0.31]	^**
Openness	–0.09 (0.09)	[–0.26, 0.08]		–0.06 (0.09)	[–0.23, 0.11]		–0.03 (0.09)	[–0.14, 0.20]
Emot. stab.	–0.02 (0.08)	[–0.17, 0.13]		0.14 (0.07)	[–0.01, 0.28]		–0.16 (0.08)	[–0.30, –0.01]	^*
Hiring decision	–0.08 (0.04)	[–0.16, 0.01]		–0.07 (0.04)	[–0.15, 0.01]		–0.01 (0.05)	[–0.10, 0.08]

Significance values are < 0.001 if ^***;< 0.01 if ^**; and less than 0.05 if ^*aWhere other co-efficients indicate the difference in mean ratings, this co-efficient indicates the marginal effect on the probability of a positive outcome. The p-value is comparable to the logistic regression model used, but the co-efficient and se is not comparable to the odds ratio in that model.

Only three of the 15 comparisons involving the personality outcomes were statistically significant, but the direction of these differences did not align with prior hypotheses and our research findings on general population samples. Those receiving unemployment benefits were rated more Agreeable than the unemployed and employed characters (zs > 2), and unemployed characters were rated as more Emotionally Stable (zs > 2) than benefit recipients. Although the correlation matrix indicated that ratings of the employment-related outcomes were more strongly associated with Conscientiousness than the other personality measures, the differences in ratings of personality between the employed and unemployed (unemployed and benefit recipient) characters were weaker than observed in previous studies [40 –42] and not statistically significant (zs < 1.5). Fig. 1 graphically depicts the results across the three experimental conditions for the four employment-related outcomes.

Fig. 1

Main effect of condition on marginal mean ratings for calibre, employability, soft skills and main effect of condition on the probability for hiring decision.

In addition to the main effect of employment-benefit status, the adjusted models in Table S1 (Appendix) show a strong effect for volunteering; characters who volunteered were rated significantly higher on all employment-related and personality outcomes except for Openness compared to those who did not volunteer. There were no interaction effects for any of the covariates by experimental condition indicating that the additional information provided about the characters did not systematically alter perceptions tied to the employment and welfare status.

4 Discussion

The broad aim of the current study was to examine the roles of unemployment stigma and welfare stigma separately on recruiters’ perceptions of job applicants and on their likely recruitment decisions. The study used a vignette rating task to assess whether a character’s employment status or benefit status was associated with perceptions of the character’s personality and with employment-related outcomes. The analysis yielded evidence of unemployment stigma (consistent with general population studies) but found no consistent difference in perceptions of those described as unemployed or unemployed and receiving benefits in a sample of Australian recruiters. This is inconsistent with previous research drawing on general population samples. The main results showed that employed characters were perceived more positively on measures of employability compared to characters who were not employed (i.e., those in either the unemployed or the unemployed benefit condition). The employed characters in the vignettes were rated as higher ‘calibre’ workers, more employable and were considered to possess better ‘soft skills’ (e.g., organisational and team work skills) Contrary to expectations, there was little evidence that unemployment or benefit receipt was associated with differences in perceptions of personality compared to employed characters. The lack of difference on the measure of Conscientiousness was most surprising and differs from previous studies using general population samples which have consistently found employed characters were rated higher on Conscientiousness than unemployed/benefit recipients [40 –42].

We found a strong correlation among the three employment-related measures (calibre, employability and soft skills), the binary hiring decision, and the personality dimension of Conscientiousness. This was consistent with our expectations that these measures would all be associated with the employment and benefit status of the vignette character, based on literature showing that Conscientiousness is strongly associated with employment and career success [53, 54]. The remaining study findings, however, showed Conscientiousness differed from the employment-related outcomes. The hiring decision question was included in the study to increase similarity with an actual recruitment process where a binary yes-no decision needs to be made, although the study was not sufficiently powered to detect a difference of this magnitude for a binary outcome. While there was no statistically significant difference in our analysis of the binary hiring decision, the measure is very highly correlated with employability and calibre ratings (rhos > .7), and as a binary outcome suffers from lower power than the continuous measures. As such it is worth noting that the seven percentage point difference in the reported probability of recruiting the employed vs the unemployed characters is not inconsistent with the effects observed in the other outcomes.

Previous studies have shown that the general public differentiates between employed and unemployed/welfare recipient characters along personality dimensions: most notably on Conscientiousness, but also on Extraversion, Openness, Agreeableness and Emotional Stability [40 –42]. There was little evidence that the recruitment professionals in the current study did so. The few differences observed in Agreeableness and Emotional Stability, were inconsistent with expectations and in the opposite direction to the previous research. The employment-focus of the experimental procedures, including using study participants who make hiring decisions professionally, may be why we found no effect of negative stereotypes on perceptions of personality. This lack of effect suggests that personality as a construct is not relevant in recruitment decisions, or that professional capacity of the participants makes them more focused on employment-relevant skills, rather than other characteristics. However, employed characters were rated more positively on all employment-related outcomes compared to unemployed characters despite the fact that all other characteristics (e.g., education, experience, appearance) were balanced. Thus, those who make recruitment decisions professionally do use engagement in work as an indicator of employability and job-relevant skills and thus unemployment alone is likely to represent a specific barrier to employment in a recruitment context. Future work may consider whether there is any benefit to explicitly flagging that one is unemployed but has not chosen to take up unemployment benefits.

We expected to find evidence that welfare stigma represented an additional barrier to the employment of people in receipt of unemployment benefits. Often studied in the context of disability [69] the co-occurring stigma related to being out of work and receiving government assistance can act as a barrier to work, making a return to work after a period of unemployment difficult [70]. We did not find evidence of welfare stigma, given the differ-ences in character assessments of the unemployed and the unemployment benefit recipients were not statistically significant. While the mean scores on measures of ‘calibre’, employability, ‘soft skills’ and hiring decision showed a gradient across the three experimental conditions (employed > un-employed > unemployment benefit), the differences between the ratings of unemployed and benefit recipient characters were weak (< 0.1 of a standardised score). This could again reflect that the study was under-powered to detect such differences. Alternatively, this may also indicate that the recruitment professionals are more focused on the implications of unemployment, rather than welfare receipt, and can manage personal biases influencing their decision making.

4.1 Other relevant findings

One of the main study outcomes, ‘soft skills’, has been consistently linked to work-place success and employability, and includes communication, team work, leadership and organisational skills, often defined as the opposite of technical –or hard –skills acquired through training or a specific job [71, 72]. Soft skills are generally thought to be acquired outside training, and incorporate the ability to handle relationships with external stakeholders or clients, maintaining professional and ethical standards, being flexible, responsible, and tolerant [73]. In general, employees who possess high level of soft skills work well with others, and display a positive work attitude in a professional work setting. The current findings suggest recruiters may make inferences about soft skills of job applicants based on their current employment status, and that this perception may disadvantage those who are unemployed and/or receiving benefits. Soft skills are also associated with the personality dimension of Conscientiousness [74], although this association was not apparent in the current study.

In contrast to the negative effect of unemploy-ment, the current study also showed that voluntee-ring had an independent positive effect on all employment-related outcomes (including perceptions of soft-skills) and on perceptions of all personality dimensions apart from Openness. This important finding among those who make recruitment decisions extends earlier work by [40] using a general population sample. The magnitude of this positive volunteering effect is somewhat greater than the detriment associated with being unemployed or in receipt of benefits.

4.2 Policy and practice implications

The current study reported on a novel experimental approach to enhance understanding of factors that influence the perceptions and decisions of individuals who make hiring decisions in a professional context. There were four important aspects of the current findings: (1) There was no evidence that welfare stigma biased the recruitment decisions or perceptions of recruiters. Lack of employment, however, was associated with more negative perceptions of a person’s employability. Compared to the general population, those in professional recruitment positions seem less influenced by the negative stereotypes associated with benefit receipt but their negative perceptions are more related to unemployment of the characters, showing evidence of ‘unemployment stigma’. To lessen the impact of stereotypes in the employment context, training and education around the implicit biases impacting recruitment decisions could be routinely implemented as part of professional development activities in organisations. Additionally, organisations can review their recruitment protocols to ensure only job-relevant information is collected from applicants, although this should be done with caution: de-identification of job applications may preclude positive discrimination towards some minority groups, as was found in Australian study senior level managers in the public service [75]; (2) These results suggest that highlighting participation in volunteering activities in job applications may help compensate for the negative perceptions employers may have towards those not currently engaged in work. However, the current study showed that the benefits of current employment and volunteering were independent: the positive effects of volunteering improved perceptions of those who were both unemployed and employed. Further, volunteering was associated with a positive impact on personality and employment dimensions whereas the adverse impact of current unemployment was restricted to work-related attributes. The current results showing benefits of volunteering should not, however, be seen as evidence to mandate participation in community-based programs and organisations should develop formal mechanisms to monitor and minimise the potential risk of exploitation of those already in a vulnerable position [76]. The benefits observed (particularly on personality) may be contingent on the voluntary and self-directed nature of the activity. In addition, evaluation of the Australian “Work for the Dole” program showed the program was associated with poorer employment outcomes, likely reflecting reduced opportunity for job-search activity [77]. Extension of the current methodology may provide further insights into this important policy area; (3) Whether because the study was restricted to individuals professionally involved in making recruitment decisions or the recruitment-focus of the task, we found little evidence that unemployment and benefit receipt influenced perceptions of personality. This stands in contrast to findings from general vignette perception tasks drawing on the general population; and, finally (4) the stigmatisation of welfare recipients and those not in employment is considered a policy tool by some to drive individuals into work, even when individuals are unable to work due to legitimate reasons and need the support [7 –10]. However, if people who are receiving benefits internalise these stigmatising beliefs or if employers act on them, the original aims of the policy will be undermined. This may be particularly damaging in the current context of high unemployment and benefit receipt around the world associated with the COVID-19 crisis.

4.3 Limitations and strengths

While the current study sought to approximate a recruitment process, it was conducted in an artificial online context, based on written vignettes. We mimicked a two-stage recruitment process (with an initial shortlisting process preceding interviews), but withheld the key experimental information on employment and benefit status until stage two and provided little variation in applicant characteristics at stage one. Extending the current methods to have greater consistency with real world employment situations, possibly including face-to-face interaction (e.g., interviews), or use of videos of interview segments should be considered. The vignette tasks did not provide specific detail about the position or type of job that was being recruited. Similarly, there was a lack of information about the background, employment context and specific recruitment experiences of those who participated in the study. Future research could improve study sensitivity by examining specific industries or occupations, and more explicitly controlling applicants’ job-relevant background. However, this would be at the expense of generalizability of the findings. While the analysis drew upon a sample of over 200 recruitment professionals, some of the key experimental comparisons (e.g., the binary recruitment decision) lacked power and future research should consider larger sample sizes. As expected, the measures of ‘employability’ were highly correlated and should be considered as measuring the same construct. Because employability was a major study focus, we pre-registered multiple measures using different wording and response scales to be able to comprehensibly examine this important aspect of the study: while the two questions about worker ‘calibre’ and the degree to which the participants perceived the person as ‘employable’ were both negatively associated with unemployment status (no reference to welfare benefit), the actual hiring decision was not. Finally, the current study methods provided much more information than is usually provided in previous studies of stereotypes, largely so the task better reflected a recruitment process. However, this additional detail may reduce the likelihood that study participants would draw on stereotypes as a basis for their decision making process [39]. Despite these limitations, the study strengths should also be recognised. The design and methodology, using two step multi-character randomisation and varying multiple pieces of information replicated a recruitment process. The experimental vignette design, manipulating multiple characteristics, masked the experimental purpose and extended past work of the authors using general population samples.

Footnotes

Acknowledgments

This paper was funded by the Australian Research Council (ARC) grant # DP16014178.

Conflict of interest

None to report.

References

Taylor

, Saunders

, editors. The price of prosperity: The economic and social costs of unemployment. UNSW Press; 2002.

Gilbert

, Besharov

. Welfare states amid economic turmoil: adjusting work-oriented policy. Policy Polit. 20111;39(3):295–308.

Department of Social Services. Final Report –A New System for Better Employment and Social Outcomes. Canberra, Australia: Author. 2015

Danziger

, Seefeldt

. Barriers to employment and the “hard to serve”: Implications for services, sanctions, and time limits. Focus. 2002;22(1):76–81.

Davidson

. Did ‘Work First’ Work?: The Role of Employment Assistance Programs in Reducing Long-term Unemployment in Australia (1990-2008). Aus Bull Labour. 2011;37(1):51–96.

Vahid Shahidi

, Siddiqi

, Muntaner

. Does social policy moderate the impact of unemployment on health? A multilevel analysis of 23 welfare states. Eur J Public Health. 2016;26(6):1017–22.

de Vries

. Negative attitudes towards welfare claimants: The importance of unconscious bias. In The Social Legitimacy of Targeted Welfare. 2017. Edward Elgar Publishing.

Blalock

, Tiller

, Monroe

. “They Get You Out of Courage:” Persistent Deep Poverty Among Former Welfare-Reliant Women. Fam Relat. 2004;53(2):127–37.

Baumberg

. Three Ways to Defend Welfare in Britain. J Poverty Soc Justice. 2012;20(2):149–61.

10.

Contini

, Richiardi

. Reconsidering the effect of welfare stigma on unemployment. J Econ Behav Organ. 2012;84(1):229–44.

11.

Saunders

, eds. Welfare to work in practice: social security and participation in economic and social life. Taylor & Francis; 2017.

12.

Schofield

, Butterworth

. Patterns of welfare attitudes in the Australian population. PLoS One. 2015;10(11):e0142792.

13.

Stuber

, Schlesinger

. Sources of stigma for means-tested government programs. Soc Sci Med. 2006;63(4):933–45.

14.

Fiske

. Stereotype content: Warmth and competence endure. Curr Dir Psychol Sci. 2018;27(2):67–73.

15.

Mood

. Take-Up Down Under: Hits and misses of means-tested benefits in Australia. Eur Sociol Rev. 2006;22(4):443–58.

16.

Riach

, Rich

. Field experiments of discrimination in the market place. The Econ J. 2002;112(483):F480–518.

17.

Wozniak

, MacNeill

. Racial discrimination in the lab: Evidence of statistical and taste-based discrimination. J Behav Exp Econ. 2020;85:101512.

18.

Koch

, D’Mello

, Sackett

. A meta-analysis of gender stereotypes and bias in experimental simulations of employment decision making. J Appl Psychol. 2015;100(1):128.

19.

Mölders

, Brosi

, Bekk

, Spörrle

, Welpe

. Support for quotas for women in leadership: The influence of gender stereotypes. Hum Resour Manag. 2018;57(4):869–82.

20.

Blanton

, Jaccard

, Strauts

, Mitchell

, Tetlock

. Toward a meaningful metric of implicit prejudice. J Appl Psychol. 2015;100(5):1468.

21.

Goodwin

, Gubin

, Fiske

, Yzerbyt

. Power can bias impression processes: Stereotyping subordinates by default and by design. Group Process Intergroup Relat. 2000;3(3):227–56.

22.

Macrae

, Bodenhausen

, Milne

, Jetten

. Out of mind but back in sight: Stereotypes on the rebound. J Pers Soc Psychol. 1994;67(5):808.

23.

Eriksson

, Rooth

. Do employers use unemployment as a sorting criterion when hiring? Evidence from a field experiment. Am Econ Rev. 2014;104(3):1014–39.

24.

Kroft

, Lange

, Matthew Notowidigdo

. Duration Dependence and Labor Market Conditions: Evidence from a Field Experiment. Q J Econ. 2013;128(3):1123–67.

25.

Atitsogbe

, Mama

, Sovet

, Pari

, Rossier

. Perceived employability and entrepreneurial intentions across university students and job seekers in Togo: The effect of career adaptability and self-efficacy. Front Psychol. 2019;10:180.

26.

De Battisti

, Gilardi

, Guglielmetti

, Siletti

. Perceived employability and reemployment: Do job search strategies and psychological distress matter? J Occup Organ Psychol. 2016;89(4):813–33.

27.

, Wang

, Lee

, Lin

, Guo

. How Machiavellianism, Psychopathy, and Narcissism Affect Sustainable Entrepreneurial Orientation: The Moderating Effect of Psychological Resilience. Front Psychol. 2019;10:779.

28.

Yeves

, Bargsted

, Cortes

, Merino

, Cavada

. Age and perceived employability as moderators of job insecurity and job satisfaction: A moderated moderation model. Front Psychol. 2019;10:799.

29.

Lowden

, Hall

, Elliot

, Lewin

. Employers’ perceptions of the employability skills of new graduates. London: Edge Foundation. 201126.

30.

Saad

, Majid

. Employers’ perceptions of important employability skills required from Malaysian engineering and information and communication technology (ICT) graduates. Global Journal of Engineering Education. 2014;16(3):110–5.

31.

Landy

. Stereotypes, bias, and personnel decisions: Strange and stranger. Ind Organ Psychol. 2008;1(4):379–92.

32.

Arai

, Bursell

, Nekby

. The reverse gender gap in ethnic discrimination: Employer stereotypes of men and women with Arabic names. Int Migr Rev. 2016;50(2):385–412.

33.

McKay

. Benefits, poverty and social justice. J Poverty Soc Justice. 2014;22(1):3–10.

34.

Stuber

, Kronebusch

. Stigma and other determinants of participation in TANF and Medicaid. J Policy Anal Manage. 2004;23(3):509–30.

35.

Stuber

, Schlesinger

. The sources of stigma in government means-tested programs. Soc Sci Med. 2006;63(4):933–45.

36.

Larsen

. The institutional logic of welfare attitudes: How welfare regimes influence public support. Comp Political Stud. 2008;41(2):145–68.

37.

Van Oorschot

, Meuleman

. Popular deservingness of the unemployed in the context of welfare state policies, economic conditions and cultural climate. In How welfare states shape the democratic public. 2014, 31. Edward Elgar Publishing.

38.

Aarøe

, & Petersen

M.B

. Crowding out culture: Scandinavians and Americans agree on social welfare in the face of deservingness cues. J Pol. 2014;76(3):684–697.

39.

Aarøe

, Petersen

. Crowding out culture: Scandinavians and Americans agree on social welfare in the face of deservingness cues. J Pol. 2014;76(3):684–97.

40.

Schofield

, Butterworth

. Are negative community attitudes toward welfare recipients associated with unemployment? evidence from an australian cross-sectional sample and longitudinal cohort. Soc Psychol Pers Sci. 2018;9(5):503–15.

41.

Schofield

, Haslam

, Butterworth

. The persistence of welfare stigma: Does the passing of time and subsequent employment moderate the negative perceptions associated with unemployment benefit receipt? J Appl Soc Psychol. 2019;49(9):563–74.

42.

Suomi

, Schofield

, Butterworth

. Unemployment, Employability and COVID How the Global Socioeconomic Shock Challenged Negative Perceptions Toward the Less Fortunate in the Australian Context. Front Psychol. 2020;11:2745.

43.

Chan

, McCrae

, De Fruyt

, Jussim

, Löckenhoff

, De Bolle

, Costa

Jr , Sutin

, Realo

, Allik

, Nakazato

. Stereotypes of age differences in personality traits: Universal and accurate? J Pers Soc Psychol. 2012;103(6):1050.

44.

Goldberg

. The development of markers for the Big-Five factor structure. Psychol Assess. 1992;4(1):26.

45.

Hogan

, Hogan

, Roberts

. Personality measurement and employment decisions: Questions and answers. Am Psychol. 1996;51(5):469.

46.

Bardi

, Goodwin

. The dual route to value change: Individual processes and cultural moderators. J Cross Cult Psychol. 2011;42(2):271–87.

47.

Saucier

, Goldberg

. The language of personality: Lexical perspectives. The five-factor model of personality: Theoretical perspectives. 1996:21–50.

48.

Digman

. Higher-order factors of the Big Five. J Pers Soc Psychol. 1997;73(6):1246.

49.

Srivastava

. The five-factor model describes the structure of social perceptions. Psychol Inq. 2010;21(1):69–75.

50.

Cuddy

, Fiske

, Glick

. Warmth and competence as universal dimensions of social perception: The stereotype content model and the BIAS map. Adv Exp Soc Psychol. 2008;40:61–149.

51.

Ward

, Thorn

, Clements

, Dixon

, Sanford

. Measurement of agency, communion, and emotional vulnerability with the Personal Attributes Questionnaire. J Pers Assess. 2006;86(2):206–16.

52.

Abele

, Hauke

, Peters

, Louvet

, Szymkow

, Duan

. Facets of the fundamental content dimensions: Agency with competence and assertiveness—Communion with warmth and morality. Front Psychol. 2016;7:1810.

53.

Barrick

, Mount

, Judge

. Personality and performance at the beginning of the new millennium: What do we know and where do we go next? Int J Sel Assess. 2001;9(1-2):9–30.

54.

Cuesta

, Budría

. Unemployment persistence: How important are non-cognitive skills? J Behav Exp Econ. 2017;69:29–37.

55.

, Shih

, Walters

, Pittinsky

. The Stigma of Unemployment: When joblessness leads to being jobless. 2011. IRLE working papers, University of California.

56.

Pearn

. Unemployment Stigma and the Selection Process. 2015. (Doctoral dissertation, Middle Tennessee State University).

57.

Karpinska

, Henkens

, Schippers

. The recruitment of early retirees: a vignette study of the factors that affect managers’ decisions. Ageing Soc. 2011;31(4):570–89.

58.

Mulders

, van Dalen

, Henkens

, Schippers

. How likely are employers to rehire older workers after mandatory retirement? A vignette study among managers. Economist (Leiden). 2014;162(4):415–31.

59.

Kervyn

, Fiske

, Yzerbyt

. Integrating the stereotype content model (warmth and competence) and the Osgood semantic differential (evaluation, potency, and activity). Eur J Soc Psychol. 2013;43(7):673–81.

60.

Roberts

, Chernyshenko

, Stark

, Goldberg

. The structure of conscientiousness: An empirical investigation based on seven major personality questionnaires. Pers Psychol. 2005;58(1):103–39.

61.

Gosling

, Rentfrow

, Swann

Jr . A very brief measure of the Big-Five personality domains. J Res Pers. 2003;37(6):504–28.

62.

Saucier

. Mini-Markers: A brief version of Goldberg’s unipolar Big-Five markers. J Pers Asses. 1994;63(3):506–16.

63.

Porzak

. Multidimensional adaptive diagnosis in career planning. Multidimensional adaptive diagnosis in career planning. Construction and psychometric properties of Career Tree inventories. 2018; Innovatio Press Wyższej Szkoły Ekonomii iInnowacji. Poland.

64.

Remedios

. The role of soft skills in employability. Int J Manag Rev. 2012;2(7):1285.

65.

Brown-Iannuzzi

, Dotsch

, Cooley

, Payne

. The relationship between mental representations of welfare recipients and attitudes toward welfare. Psychol Sci. 2017;28(1):92–103.

66.

Dyck

, Hussey

. The end of welfare as we know it? Durable attitudes in a changing information environment. Public Opin Q. 2008;72(4):589–618.

67.

Hedeker

, Gibbons

, Flay

. Random-effects regression models for clustered data with an example from smoking prevention research. J Consult Clin Psychol. 1994;62(4):757.

68.

Cameron

, Trivedi

. Microeconometrics using stata. College Station, TX: Stata press; 2009.

69.

Friedman

. The relationship between disability prejudice and disability employment rates. Work. 2020; (Preprint): 1-8.

70.

Nilsson

, Ekberg

. Employability and work ability: Returning to the labour market after long-term absence. Work. 2013;44(4):449–57.

71.

Litecky

, Arnett

, Prabhakar

. The paradox of soft skills versus technical skills in IS hiring. J Comput Inf Syst. 2004;45(1):69–76.

72.

Canovi

, Succi

. Be Yourself: How Role Identities Influence Transgenerational Entrepreneurship in Family Firms. In Academy of Management Proceedings 2019 (Vol. 2019, No. 1, p. 13285). Briarcliff Manor, NY 10510: Academy of Management.

73.

Kim

, Erdem

, Byun

, Jeong

. Training soft skills via e-learning: international chain hotels. Int J Contemp Hosp Manag. 2011.

74.

Hamid

, Mohamed

, Devarajoo

, Mani

, Lazim

. The Relationship between Big Five Personality and Soft Skills among Graduate Unemployment. Int J Psychosoc Rehabilitation. 2020;24(2).

75.

Hiscox

, Oliver

, Ridgway

, Arcos-Holzinger

, Warren

, & Willis

. Going blind to see more clearly: Unconscious bias in Australian Public Service shortlisting processes. 2017. Behavioural Economics Team of the Australian Government.

76.

Krinsky

, Simonet

. Servitude and volunteer work: The political exploitation of invisible labor in New York City parks. Soc Contemp. 2012;3:49–74.

77.

Borland

, Tseng

. Does ‘Work for the Dole’ work?: an Australian perspective on work experience programmes. Appl Econ. 2011;43(28):4353–68.