Can we make fair algorithms?

In some ways, artificial intelligence acts like a mirror. Machine learning tools are designed to detect patterns, and they often reflect back the same biases we already know exist in our culture. Algorithms can be sexist, racist, and perpetuate other structural inequalities found in society. But unlike humans, algorithms aren’t under any obligation to explain…

Uber is an economics research powerhouse

Like Yahoo and Google, Uber offers access to the kind of data that, as recently as 10 years ago, an economist could only dream of. Uber’s more than 3 million drivers provide roughly 15 million trips, globally, every day. Uber’s researchers can test vital questions about driver pay, customer satisfaction, and urban transit with tiny…

Amazon’s AI recruiting tool discriminated against women

The explanation for what happened seems odd; a preponderance of males in a field would not necessarily result in an algorithm that discriminates against women. This is the problem with machine learning; it really can be a black box. Amazon.com Inc’s (AMZN.O) machine-learning specialists uncovered a big problem: their new recruiting engine did not like…

IES proposes Standards for Excellence for ed research studies

I predict pre-registration will make some people very unhappy. Register Studies. Did the researcher execute the research and analysis activities as originally proposed in a recognized study registry? Did the registration describe key elements of the study protocol, including a limited number of primary outcomes?  Are any deviations from those plans clearly documented and their rationale…

Sociology journals did not fall for the dog rape culture hoax

Most of the news coverage missed this: There are many fields of academia that have absolutely no patience for nonsense. While the hoaxers did manage to place articles in some of the most influential academic journals in the cluster of fields that focus on dealing with issues of race, gender, and identity, they have not…

Contrary to common belief, memories are just not dependable

This was one of my main points in critiquing college student surveys: we can’t ask students questions as if they have a computer hard drive in their head. A similar study conducted in 1988 yielded similar results. Researchers showed participants slides depicting a burglar stealing a hammer from someone’s office. Then they read a brief…

You know those presenters who always go over time?

It’s not just your imagination boring speakers drone on. At least according to a small study reported in a letter to Nature: “I investigated this idea at a meeting where speakers were given 12-minute slots. I sat in on 50 talks for which I recorded the start and end time. I decided whether the talk…

HARKing: Hypothesizing after the results are known

This article considers a practice in scientific communication termed HARKing (Hypothesizing After the Results are Known). HARKing is defined as presenting a post hoc hypothesis (i.e., one based on or informed by one’s results) in one’s research report as if it were, in fact, an a priori hypotheses. Several forms of HARKing are identified and…