Feedback as a strategy for overcoming automatic gender stereotypes

Eimear Finnegan, Jane Oakhill, Alan Garnham

Research output: Contribution to conferencePosterpeer-review


Stereotypical gender information is automatically evoked when certain role names and professional terms are read. As a consequence, people expect surgeons to be male and nurses to be female. Such inferences are difficult to suppress and result in processing difficulty when violations of stereotypical gender are encountered. In a series of three studies, we investigated feedback as a form of stereotype reduction training. We hypothesised the feedback would create an awareness of stereotyping and facilitate control over the immediate activation of gender stereotypes. Specifically, we compared the value of individual performance-related feedback versus group-consensus feedback in lowering gender stereotype biases.
The basic task was a judgement task devised by Oakhill, Garnham and Reynolds (2005). Across 3 blocks of trials participants were presented with word pairs comprised of a role name with stereotypical gender and a kinship term with
definitional gender e.g. Surgeon/Mother. The task was to decide as quickly as possible whether both terms could refer to one person. Importantly, the three experiments differed only in Block 2 where the type of feedback was manipulated.
In Experiment 1 feedback was based on the participants’ performance, simply indicating whether they were ‘Correct’ or ‘Incorrect’ after each response. In Experiment 2 feedback consisted of a statement informing participants of their
accuracy and the (fictitious) percentage of participants in a previous experiment that agreed or disagreed with their choice. If participants deemed gender incongruent trials to be incorrect (thus succumbing to stereotype biases) they
would receive feedback indicating that a very high percentage of previous participants disagreed with their response thereby highlighting their comparatively strong personal tendency to stereotype. In a third, control, experiment no feedback was given.
Stereotype activation on gender incongruent word pairs was most succesfully reduced in Experiment 1 with accuracy increasing significantly across all 3 blocks (+14.5% from Block 1 to Block 3). In Experiment 2 a smaller but significant improvement was also found (+9.7%).However in the Control experiment there was no significant improvement across blocks (+0.83%). Results suggest the use of individual performance related feedback is highly effective in reducing the gender stereotyping effect in this judgement task. The positive influence of group-consensus feedback should also be noted.
The reaction time data showed a similar pattern with the biggest decrease in reaction times to gender incongruent word pairs (– 412ms from Block 1 to Block 3), followed by Experiment 2 (– 308ms). However, there was also a significant
improvement in the Control experiment (– 238 ms) suggesting that faster reaction times across blocks in each of the experiments were due, at least to some extent, to practice effects.
Our findings provide further support for the malleability of stereotype biases. Creating awareness of stereotype tendencies through providing feedback on behaviour is a valuable method of reducing activation of gender stereotypes, though further research is needed to investigate what cognitive or motivational changes underlie such effects, and to determine whether such training can produce longer term reductions in stereotyping.
Original languageEnglish
Publication statusPublished - 8 Oct 2012
EventArchitectures and Mechanisms for Language Processing - Riva del Garda, Italy
Duration: 6 Oct 20128 Oct 2012


ConferenceArchitectures and Mechanisms for Language Processing
CityRiva del Garda


  • judgement task
  • stereotypes
  • gender

Cite this