This is a Medium Rare template. Explore our complete range of designs here.
top
Mirroring changes in public opinion most organisations are concerned about social justice and rooting out discrimination in the workplace.
Discrimination means treating someone differently because of their age, gender, race, religion, or other protected characteristic.
An example of direct discrimination in the workplace is not offering a job to a candidate because of their skin colour.
Direct discrimination is usually the behavioural manifestation of bias or prejudice.
Organisations committed to evidence-based practice need to know which initiatives are most effective at producing meaningful change.
However finding evidence for which interventions effectively reduce discrimination in the workplace is difficult.
While there are numerous studies of interventions designed to reduce discrimination many only assess general, nonwork-related attitudes and behaviours.
Another difficulty is that evidence can be mixed. Some aspect of bias might be effectively reduced by an intervention while no effect is found for another aspect of bias.
A recent study by Elaine Costa published in the Journal of Applied Psychology aims to address these problems.
Costa identified 70 peer-reviewed papers covering interventions involving workplace outcomes and intergroup bias producing 208 intervention effect sizes for analysis.
Costa started the review by categorising interventions based on the type of bias they target and whether they are active or passive.
Inspiration for Costa’s categories comes from a general model of attitudes which has three dimensions. The dimensions of this model are:
In the context of attitudes about other groups the cognitive dimension maps to stereotypes, the affective dimensions maps to prejudice and the behavioural dimension maps to discrimination.
Costa defines active interventions as those which directly target attitude dimensions.
Passive interventions do not attempt to change any specific attitude dimension and rely on the individual to be motivated to update their biases and behaviours.
Costa came up with the following bias reduction intervention categories.
Interventions aimed at disrupting stereotype processing target the cognitive component of attitudes. These approaches work to interrupt the activation or use of stereotypes, influencing the type of information an individual prioritizes.
Examples include anonymizing CVs and structured evaluation processes.
Interventions designed to update affective states primarily address the affective component of attitudes. These strategies aim to enhance positive emotions or reduce negative feelings toward the person being evaluated.
Examples include imagining positive interactions or engaging in perspective-taking exercises.
Interventions aimed at inhibiting bias manifestation focus on addressing the behavioural component of attitudes. These approaches create strong external incentives to discourage individuals from acting on stereotypes or biases.
Examples include measures like accountability systems and affirmative action policies.
Interventions focused on educating about bias processes aim to raise awareness, provide information, or serve as reminders about biases.
Examples include initiatives such as equality messaging and diversity training programmes.
Categorising the interventions helps to answer the question which intervention categories most effectively reduce which types of biased outcomes - stereotypes, prejudices and discrimination.
Costa hypothesised that “interventions targeting the behavioural, affective, and cognitive dimensions of attitudes produce better outcomes on measures of workplace discrimination, prejudice, and stereotyping when there is alignment between the targeted and the measured dimensions.”
Costa also predicted that educating about bias processes would be less effective at reducing bias outcomes compared to the other intervention categories.
Another aim of the study was to identify which intervention types are most effective in reducing the behavioural component discrimination in the workplace.
Costa’s analysis found that when interventions are aligned with outcomes the average effect size is moderate compared to a small effect size when there is no alignment.
The analysis also found that educating about bias processes is less effective than other interventions.
On the question of which intervention types are most effective at reducing discrimination accountability was found to be the best overall.
Accountability is in the inhibiting bias manifestation category. It requires evaluators to explain or justify their ratings, decisions or conclusions.
Of the interventions in the disrupting stereotype processing category structured evaluations were most the most effective.
Structured evaluations use consistent criteria and anchored rating scales, focusing on clear benchmarks to minimize ambiguity and reduce bias from stereotypes or heuristics.
In the updating affective states category, the most effective intervention was imagining contact which encourages evaluators to envision a positive interaction with a target group member before evaluation, fostering empathy and reducing prejudice.
