Table 1.
Data description.
Fig 1.
Number and percentage of female faculty members by field of research.
Overall, 32% of the faculties in the data are female, and the percentage within a bar is the percentage of female faculties within the research field.
Fig 2.
Estimated annual number of patent applications and patents granted per person by field of research.
Left: annual number of patent applications per person; right: annual number of patents granted per person. The annual number of patents is estimated by Poisson regression, where the number of patents is the response variable and the field of research and the time since graduation are the independent variables.
Fig 3.
Estimated annual number of patent applications and patents granted per person by gender and time of graduation.
Left: annual number of patent applications per person; right: annual number of patents granted per person. The annual number of patents is estimated by Poisson regression, where the number of patents is the response variable and the interaction of gender by time of study is the independent variable.
Table 2.
Poisson regression results on the annual number of patent applications per person.
Table 3.
Poisson regression results of annual number of granted patents per person.
Fig 4.
Estimated ratio of annual number of patent applications and granted patents per person between male and female faculties with 95% CIs.
Left: patent applications; right: patents granted. The ratios are estimated using generalized linear models with Poisson distributions. The number of patent applications within the given time period is the response variable, and university, title, field, and gender within the given time period are independent variables.
Fig 5.
Estimated probability of a faculty being the first inventor of a patent.
Left: patent applications; right: granted patents. The probabilities are estimated using generalized linear models with binomial distributions. The number of patents and the number of patents where the faculty is the first inventor together are the response variables, and group, gender, and the interaction between group and gender are the independent variables.
Table 4.
Generalized regression results of probability of being the first inventor in a patent application.
Table 5.
Generalized regression results of probability of being the first inventor in a granted application.
Fig 6.
Estimated odds ratio of being the first inventor between female and male faculties with 95% CIs.
Left: patent applications; Right: granted patents. Odds ratios are estimated using generalized linear models with binomial distributions. The number of patents and the number of patents in which the faculty member is the first inventor combined are the response variables. University, title, field, gender, group, and the interaction of gender and group are the independent variables.
Fig 7.
Kaplan-Meier estimates of time to first patent after graduation by gender.
Left: Time to first patent application; Right: Time to first patent grant.
Table 6.
Cox proportional hazards regression results of time to first patent application.
Table 7.
Cox proportional hazards regression results of time to first patent application.