We know that people use status characteristics (widely-held societal expectations about a group) as a shortcut to understanding how a group (e.g. women, men) might perform and how valuable the group is. Sociological research has shown us that gender is an important status marker–male gender is seen as higher status while female gender is seen as lower status. This study uses computers named Julie and James to test how people apply the status marker of “gender” while controlling for all other factors. In particular, the researchers observed the different ranking that occurred when people evaluated the performance of the computers named James and Julie. Ultimately, participants estimated the value of James-computers to be much higher than that of the Julie-computers.
Prior research has shown that people’s perceptions of their technology-enabled devices can be shaped by social categories. This paper applies the status characteristics theory (SCT) from sociology to interactions between a human and a computer. Status characteristics are human attributes that influence people’s expectations of performance and status processes (ranking) emerge when these expectations of performance vary by group (e.g. women, men). We might think this is only applicable to narrow factors like mathematical ability, but it applies more broadly to diffuse characteristics like intelligence.
This paper sought to determine when status characteristics (e.g. performance, perceived economic value) become relevant in human-computer interactions and what status processes (ranking) emerge.
We assume that certain characteristics are universal in groups unless individuals prove otherwise or somehow disassociate the characteristic from the task upon which they are being evaluated (this is known as the “burden-of-proof” assumption–where it is left to the individual to prove otherwise). It is well-documented that traits such as sexual orientation, race, ethnicity, and gender have influenced the way people perceive performance on tasks unrelated to these social categories.
This study observed users’ expectations, ratings, and valuations of identical computers named James and Julie and revealed three important findings. Men and women alike (in fact, many more women participated in the study than men) had equal confidence in the James and Julie computers; and evaluated the machines’ performance equally.
The machines were identical in performance but users assigned a monetary value to Julie-computers that was 25% less than James-computers.
However, even with no perceived performance difference, the economic value applied to James and Julie computers was significantly different. The machines were identical in performance but users assigned a monetary value to Julie-computers that was 25% less than James-computers.
Gender inequality is well-documented in the labor market in regard to salaries, occupations, and positions of leadership. Given this pervasive inequality in the economic marketplace, this study suggests that the mere reference of economic value may activate people’s assumptions about women’s performance (expected to be less) compared to men’s (expected to be higher).
- Recognize underlying biases – These results are consistent with other studies that have shown how equally performing humans are still devalued because of gender characteristics. This provides additional evidence that we need to be gender-blind or gender aware when we make evaluations.
- Be aware of unconscious biases during the development of a product – Gender is relevant to how people value technology and this may also extend to the implicit gendering of products. As we think about personifying technologies, care should be taken when naming or developing characteristics of technology products.