Followup Experiments for "What Jobs Come to Mind?"

We conduct several different survey experiment exercises to explore the mechanisms and robustness of the main stereotyping result in our paper "What Jobs Come to Mind?".

We have three main surveys.
1. IAT: We conduct an implicit association test where participants must sort words (which are either occupations or fields-of-study) into groups of majors and careers (pulled from our previous paper). Each participant undergoes 4 rounds, where each round has a focal career and major group. These pairs of career and major groups include each of our ten major groups paired with teaching and business, plus (where necessary) each major group paired with its most distinctive career group. Participants then sort words into the focal major vs other majors and/or the focal career vs other careers. The key variable is the difference between how quickly participants correctly sort words into "Focal Major or Focal Career" vs "Other Majors or Other Careers" minus how quickly they sort them into "Focal major or Other careers" vs "Focal career vs Other Majors", and how this differs by whether the career-major pair are stereotypical or not.

Note that this design includes within person randomization in the timing of these rounds. The "control" condition (major + career sorting aligned in the 4th or 5th round) and the "mixed" condition (unaligned, occurring in the 5th or 4th round) are in a randomized order at the end of each IAT round. We can thus look at the effect of these conditions on time required to sort, netting out any order effects.

2. IAT + Beliefs: This survey contains our main "population beliefs" questions (i.e., about the share of Americans with major M who work in each career c) along with the IAT exercise described above. The only difference between the IAT in this survey and the IAT described in (1) above is that here we will only include stereotypical career-major pairs. We will explore heterogeneity in population beliefs questions based on the IAT performance.
3. Other Measures: This survey contains i) our main "population beliefs" questions, ii) a survey eliciting which careers participants are "aware" of, iii) a direct elicitation from participants of the sorts of jobs that come to mind when they consider their stereotype of a randomly selected major, and iv) an exercise that asks participants to decide which of two careers they more associate a given major with (repeated many times for difference careers/majors). We will explore heterogeneity in the population beliefs questions by these other measures.

In all surveys, we will additionally collect respondents' own occupation and major (if applicable), which we will compare to participants' population beliefs.

2025-05-07

2025-05-21

Our primary outcomes are participants population beliefs, their IAT times, measures of which careers they are (un)aware of, the direct stereotype elicitation, and which careers they associate more with each major.

For timing variables, we will need to account for outliers and will do so by winsorizing the data and/or by taking medians.

We conduct a variety of robustness checks and explorations of the mechanisms behind the stereotyping result in our paper "What jobs come to mind?"

See the "Intervention" section above.

All randomization will occur within Qualtrics.

Most randomization occurs within person (e.g., which careers/majors the IAT pertains to, which jobs participants are asked their awareness of). For the direct stereotyping elicitation, we randomize across participants which major group they are asked about.

No

For the IAT survey, we plan to collect 100 responses. For the IAT + beliefs survey, we plan to collect 400 responses. For the Other Measures survey, we plan to collect 250 responses.

See above.

See above.