Does Wording-Based Big Data Have Enough Practical Value？

[khan1792]

Your paper finds that gender neutral wording does not correlate with job seekers', probably even employers', underlying bias. It means that this set of big data does not touch the essence of the problem about gender bias because many gender-bias-like problems can only be represented by action or underlying bias rather than politically correct wording (such as gender neutral posting). So I want to ask if wording-based big data is short of practical meaning to some extent, or if there are some methods improving its practical value?

[tamos]

I felt this was actually a really interesting point. It clearly demonstrates the limits of applying a "gender neutralizing" filter to job advertisements, while at the same time it shows how far things have come (overall trend of more gender neutral job posts).

I wonder if this technique has the most value as a way to evaluate/track the development of an abstract construct (~ gender equality in recruiting). If that was the case, it would be enormously useful to many actors. From my experience, a great deal of effort and funds go into attempts to approximate such concepts (in industry, government, non-profit sectors). Compared to a survey, this approach is low-cost, making it an interesting alternative.

[jgdenby]

This is an interesting question. If applicants self-select in a way that doesn't quite correspond to algorithmic evaluations of biased language, what is the use of the evaluation?

It also suggests that biased language is more of a consequence than a cause of ingrained preconceptions about job suitability.

[WZheng-94]

As I see, the gender bias in job posts reflects the intention of employers but not the employees or applicants, which means, the discrimination against females in job recruitment has been actually decreasing, at least, regarding jobs in the entry level. But whether people's conception or fundamental mechanism of gender repression has changed or radically transformed is another story. Following studies in extra fields are needed.

[khan1792]

@ZhengErWei I don’t think neutral words in posts necessarily indicate the intention of employers. I know some cases that some Chinese academic Institutions posted gender neutral job information but the employers did say they wanted to hire male employees in private occasions. They clearly understood that gender biased words must be avoided in academic posting since these words are not politically correct. As many sociological literature demonstrates, gender, race and class discrimination in contemporary society has a more indirect and publicly invisible form. So, it seems difficult to tell the authentic intention of employers if we only rely on wording-based data.

[WZheng-94]

Yaa, I agree with you on the point that the value of the research itself should have been thought over and stated more clearly in the paper. I also wonder how to distinguish the very research paper from a business report since it doesn't discuss directly the change in underlying mechanism of gender inequality or contributions to the field of social science research. But, i can only say, to some extent, the paper demonstrates some change of ideas of employers when considering hiring people, though whether the change is meaningful, or reflects the genuine shift of people's perceptions is still controversial.