Traditional admissions requirements fail to predict PhD completion in Physics

Through a rigorous statistical analysis of a sample that includes roughly one in seven students who entered physics PhD programs from 2000-2010, we find that the traditional admissions metrics of undergraduate GPA, the Graduate Records Examination (GRE) Quantitative, Verbal, and Physics Subject Tests, do not predict completion in US physics graduate programs with the efficacy assumed by admissions committees. In fact, we find only undergraduate GPA to have a statistically significant association with physics PhD completion. While we find weak correlations between the GREQ and GREP and graduate grades for some groups, these test scores are not predictive of PhD completion for any group. Noting the significant race, gender, and citizenship gaps on the GRE tests, these findings indicate that the typical physics PhD admissions process is a deterrent to expanding representation: it selects against already-underrepresented groups and US citizens with tools that fail to predict PhD completion.