Collecting demographic data in assessments is a routine yet vital practice. Such data may include information about participants such as age, gender and cultural background (i.e. known variously as sensitive or protected category data). However, clients may have concerns about the collection and use of such data.
In this article, we discuss the main reasons for collecting demographic data in assessments and why we would consider this feature of assessment projects a vital and needed step for all clients.
The two main reasons for collecting demographic data from participants during assessments are:
- Mitigation and analysis of adverse impact against protected minorities
- Continual research on and quality assurance of assessment devices and processes
Of these two reasons, we will focus mainly on the first reason, because it is the most critical from the perspective of safeguarding our clients against possible legal and reputational risks associated with incorrect application of assessment data.
Understanding adverse impact and the four-fifths rule
In most countries, fair labor practices have been established, in part, to prevent bias against employees when it comes to finding employment, being promoted, accessing company-controlled benefits and a host of similar advantages that may accrue to employees or potential employees as a result of being in the labor marketplace.
For instance, in the United States, the Equal Employment Opportunity Commission (EEOC) enforces the civil liberties laws that protect workers from discrimination and prejudice by employers.
One of the key issues relevant to assessments that concern bodies such as the EEOC is known as adverse impact.
To understand adverse impact, it is useful to look at the precedent-setting legal cases that brought it to the attention of employers, HR professionals and IO Psychologists alike.
In the Supreme Court ruling of Griggs v. Duke Power in 1971, the court ruled in favor of twelve African-American employees who claimed disparate treatment (i.e. adverse impact) when they were assessed for suitability to be employed by Duke Power.
The claimants were able to illustrate that there were systematic biases in the assessments used by Duke Power that consistently favored White candidates and disadvantaged African-American candidates.
One of the consequences of this case (and others that followed) was that employers had pay particular attention to:
- Using assessments that were closely related to job-relevant competencies and
- Removing adverse impact from their assessments that may disadvantage protected groups such as women and ethnic minorities.
How can assessment end-users establish that their practices of selection, promotion and other talent decisions do not result in adverse impact?
The answer to this question is the widely accepted “Four-Fifths Rule”.
Also advocated by the Uniform Guidelines on Employee Selection Procedures, the four-fifths rule states that when the selection rate for a certain group is less than 80 percent of that of the group with the highest selection rate, there is adverse impact on that group.
To illustrate, the following example will be helpful:
- In a given assessment project, there are 500 female candidates and 250 male candidates who apply for a position that requires pre-employment assessments.
- In this example, let’s assume that 6 female candidates were hired based on their assessment results and 5 male candidates were hired according to the same assessments.
- The selection rate for females was therefore 1.2% (6 / 500) and the selection rate for males was 2% (5 / 250).
- When we divide 1.2% by 2% as recommended by the four-fifths rule, the result is 60%, which is less than the four-fifths (80%) criteria.
Therefore, it can be assumed that women were adversely impacted by this assessment process, even though more women were selected in this instance than men.
The four-fifths rule is a robust and easy-to-use principle that companies use to ensure that their selection and assessment practices do not contain systematic biases against protected groups and is generally regarded as a necessary condition for showing a fair and equitable talent process.
In addition to the four-fifths rule, more sophisticated statistical procedures such as significance testing of group differences of assessment results may also be employed to illustrate the absence (or presence) of adverse impact, but these are often not called for if the simpler, more conservative four-fifths rule is satisfied first.
So the logical next question is: What has the collection of demographic data (especially protected category data) got to do with adverse impact?
Demographic data collection and adverse impact
The answer is quite simple:
- In order for an employer to show that their talent selection methods (either as a whole or for specific instances) are free of bias and adverse impact, the four-fifths rule will need to be applied and satisfied.
- The four-fifths rule can only be calculated if demographic data such as gender and cultural background were collected in the initial phases of talent selection and are therefore known.
This is a key principle of best-practice assessment methods. Without this data, employers are exposed to considerable risk if their talent management and selection methods are ever questioned or legally challenged.
Not only will such an oversight almost certainly bolster the case of anybody suing the employer (i.e. why would they not want to calculate whether the four-fifths rule was upheld if they were using unbiased methods?), but it also leaves the assessment end-user in the dark about the quality and even-handedness of the assessments they rely on for making decisions.
To this point, there is considerable evidence to suggest that more diverse teams tend to be more innovative, produce more profit for their employers, and are more resilient to change.
Given such findings, why would an employer not want to benefit from creating more diversity by using unbiased, fair instrumentation when selecting individuals for employment, promotion, educational opportunities and so on?
But again, ensuring the lack of adverse impact and thus ensuring more diversity can only take place if talent professionals have the needed demographic data at hand.
It may seem counterintuitive to collect protected category demographic data as a mitigation strategy for biased selection and adverse impact. On first glance, it would seem that ignoring such demographic variances is the correct attitude to adopt if one is to be equitable and fair in employing job-seekers or offering opportunities to existing employees.
But the reverse is actually true if we consider the explanations above. To show fairness, an employer needs to show that adverse impact on protected minorities is absent. And that can only be done using the objective evidence of tests such as the four-fifths rule which require collecting demographic data in assessments.
If you would like to know more about this best-practice principle of assessment, let us know at: firstname.lastname@example.org.