Perceived abilities or academic interests? Longitudinal high school science and mathematics effects on postsecondary STEM outcomes by gender and race

International Journal of STEM Education

Table 2 Dependent variable distribution in the analytic sample with multiple imputation, by gender and race (N = 11,560)

Research Questions 1–3: Predicting STEM Major Selection
	Gender X² = 438.7, p < 0.001		Race X² = 301.2, p < 0.001
	Boys	Girls	Asian	Black	Latina/o/x	Multiple/Other	White
STEM	1680	980	500	170	310	230	1440
STEM	32.0%	15.5%	42.5%	16.2%	19.5%	22.3%	21.6%
Non-STEM	3560	5340	680	890	1280	820	5230
Non-STEM	68.0%	84.5%	57.5%	83.8%	80.5%	77.7%	78.4%
Total	5240	6320	1180	1060	1590	1050	6670

Research Question 4: Predicting Computing/Engineering vs. Other Sciences
	Gender X² = 375.9, p < 0.001		Race X² = 12.8, p < 0.01
	Boys	Girls	Asian	Black	Latina/o/x	Multiple/Other	White
Computing/Engineering	1130	280	250	90	170	100	790
Computing/Engineering	67.5%	28.6%	50.4%	51.5%	56.5%	44.0%	55.0%
Social/Behavioral, Natural, and Other Sciences	550	700	250	80	140	130	650
Social/Behavioral, Natural, and Other Sciences	32.5%	71.4%	49.6%	48.6%	43.6%	56.0%	45.0%
Total	1680	980	500	170	310	230	1440

Total Number refers to the total number of observations in each gender and/or race group. The survey weight (w4w1stup1) was applied to account for nonresponse and the stratified sampling design, to enhance the external validity. The N in each subgroup was rounded to the nearest ten and the percentages rounded to the nearest tenth decimal, to comply with NCES restricted-data regulations. The total number of cases in each subgroup may not be exactly equal to that of the analytic sample because of this rounding. Statistical significances within each gender and race group were assessed using Chi-square tests, with the test statistic and p-value reported for each variable
Source: U.S. Department of Education, National Center for Education Statistics, High School Longitudinal Study of 2009 (HSLS: 09), Base Year, Student Survey, 2009; Second Follow-up, Student Survey, 2016