Redacting CVs: Is it a roll of the dice?

 

Introduction

It is now hardly a matter of controversy that in order to improve performance, organisations in both the public and private sectors need to attain a credible degree of diversity at all levels and particularly at the level of leadership. However for many organisations finding a successful means to reach an equitable degree of demographic and cognitive diversity has proved difficult.

The Blind Resume Experiment

The appraisal of CVs is a process often compromised by unconscious bias. A 2012 study by the Australian National University found that applicants with non-Anglo-sounding names had to submit many more CVs to get interviews than those with Anglo names. Similar studies have been conducted overseas (USA, UK and Canada) finding that women and minorities have a significantly lower chance of moving beyond the CV-assessment stage than Anglo males. Indeed, this result has been replicated with not just CVs but even with perceptions of the quality of work samples and reports.

One proposed method of countering this is to de-identify CVs by removing names as well as indications as to age, gender, ethnicity, social status, health and so on of the applicant.

Support for De-identification

The UK Government, the Canadian Government and the Victorian Government in Australia and a number of private organisations such as HSBC, Deloitte, Virgin Money and KPMG have all launched extensive programs to de-identify CVs as a means of levelling the playing field for diverse candidates, at least at the initial stage of CV assessment.

In Europe, field trial studies (reported by Ritte, 2012) in Germany, Sweden, the Netherlands and France, to measure the effects of anonymising resumes, disclosed clear benefits for women and minorities by getting more to the interview stage in all except France (see below).

Does de-identification help at all?

Now a new study by the Behavioural Economics Team (BETA) of the Australian Government has suggested that de-identified CVs do not necessarily help and, specifically in the case of the Australian Public Service (APS), actually prejudice women and other diverse groups. Some 2100 members of the APS across 14 departments participated in a randomised controlled trial to assess the diversity impact of de-identified CVs compared to those with names reflecting gender or ethnicity. When the CVs reflected identifiable names, females were 2.9 % more likely to be shortlisted for interviews and males 3.2% less likely. When comparing those with Anglo-sounding names to non-Anglo-sounding names, Indigenous females were 22.2% more likely to be shortlisted; Indigenous males 9.4% more likely; all minority females 8.6% more likely; all minority males 5.8% more likely and Anglo males 6.5% less likely.

The conclusions drawn by the authors are:

  • De-identification of CVs would have the unintended consequence of setting back diversity at senior management level in the APS, and
  • Senior public servants appear to be actively promoting diversity when selecting job candidates and this is a result of long–term diversity awareness raising within the APS

The authors acknowledge that there were limitations to the study in that participation was voluntary and that reviewers might behave differently in a real recruitment situation.

Some Confirmation

A paper by IZA, the Institute for the Study of Labour in Bonn entitled ”Unintended Effects of Anonymous Resumes” (2014) evaluated a scheme in which the French public employment agency offered firms a voluntary anonymous resume procedure for their recruitments. 1,613 firms were invited to participate and 1,005 (62.3%) agreed.

In this study, when de-identification of resumes occurred, “minority” applicants’ progress- to-interview rates dropped from 9.3% to 4.7% and “majority” applicants’ progress-to-interview rate increased from 11.6% to 17.7%. Thus the interview-granted gap between majority and minority candidates actually widens by 10.7% as a result of de- identifying resumes “This increase in the interview gap against minority candidates is the exact opposite effect to what policymakers intended.”[Behaghel et al.] And as a result of this study the French Government abandoned the idea of making anonymous resumes mandatory.

 No universal success for de-identification

The debate about the use of de-identified CVs as a means of furthering a diversity agenda is by no means resolved. On one hand, those who believe that the triggering of unconscious bias, when gender or racial and ethnic affiliations are revealed, is the major barrier to diverse recruitment, will favour de-identification. On the other hand, those who believe that the push to achieve significant diversity must take precedence and seen to occur at every stage of the recruitment process,will favour identification of candidates.

Furthermore, as the studies above have shown the outcomes of any particular policy are anything but predictable. Results may differ from industry to industry and even to specific organisations depending on the extent to which acceptance of diversity and inclusion principles have become embedded.

Insights from the Studies

What appears to emerge from the two studies is that de-indentification of CVs may remove critical information showing that a female or Indigenous candidate has significant potential and capability, not fully evident in their formal qualifications or work history. For example a work gap by a female may simply be a result of pregnancy and a lack of tertiary qualification in the case of an Indigenous applicant may be as result of insufficient financial resources.  Moreover if two candidates are equal in skills and potential it is impossible to prefer the diverse candidate when de-identification has occurred.

Removing important information and “neutralising” CVs may not really level the playing field. It may simply deprive diverse candidates of the opportunity to bring to the fore factors which are highly relevant and which work in their favour. And importantly the benefits of, de-identifying CVs only lasts for the initial stage of recruitment. Once the candidate moves to interview or beyond he or she may be subjected to bias unless adequate de-biasing mechanisms have been implemented throughout.

Choosing the Best Option

It is clear that de-identification of CVs can be a useful way of avoiding biased assessments by reviewers, particularly when the historical evidence reveals that particular groups such as women or ethnic minorities are being rejected in large and unjustifiable numbers. But it is not a panacea for all the potential bias that can be encountered along the recruitment journey.

As Ritte has pointed out:

“Current evidence does not seem to support the desirability of a mandatory introduction of anonymous job applications in every context. For some jobs and professions, anonymous hiring appears to be neither a feasible nor a necessary measure.”

 Decisions by management on whether to de-identify or not should depend on a large range of factors such as the nature of the industry, the type of skills being sought, the size of the potential pool of candidates and whether a culture of diversity and inclusion has been embedded in the organisation – in other words whether leaders have internalised that diversity will add value and therefore something they seek.

What is critical is that whichever route is chosen, anti-biasing measures are in place at every step of the way through recruitment and beyond as diverse employees progress on their career path.

 

Leave a Reply

Your email address will not be published. Required fields are marked *