. 2021 Sep 13;50(12):2394–2411. doi: 10.1007/s10964-021-01493-1

Table 6.

Comparisons between complete data and missing data samples within the analytic sample for this study

	Complete Data Sample			Missing Data Sample			t test or Chi-square test	Effect size
Study variables	N	M (SE)/%	Min, Max	N	M (SE)/%	Min/Max	t test or Chi-square test	Effect size
Female students	12, 710	50%	0,1	8,460	47%	0/1	86.49***	−0.06^b
Race/ethnicity	12,710	—	1,5	8,460	—	1/5	43.02***	0.05^b
Potential first-gen. college	12,710	44%	0,1	8,460	57%	0/1	198.80***	0.10^b
Family income	12,710	4.30 (0.04)	1,13	8,460	3.82 (0.09)	1/13	11.00***	0.15^a
Science identity (9^th grade)	12,710	2.32 (0.01)	1,4	8,370	2.23 (0.03)	1/4	8.69***	0.13^a
Science identity (11^th grade)	12,710	2.49 (0.01)	1,4	8,070	2.45 (0.03)	1/4	8.36***	0.12^a
Science identity (3 years post-HS)	12,710	2.62 (0.01)	1,4	7,500	2.52 (0.05)	1/4	5.03***	0.12^a
Science grade (8^th grade)	12,710	4.12 (0.02)	1,5	7,950	3.96 (0.04)	1/5	14.79***	0.24^a

Note. Frequencies displayed are weighted for both samples. ^aIndicates Cohen’s d was used for measuring effect size among independent sample t-tests for continuous variables. Standard interpretation: small effect: 0.20, moderate effect: 0.50, large effect: 0.80. ^bIndicates Cramer’s V was used for measuring effect size among Chi-square tests for dichotomous variables. Standard interpretation: small effect: 0.10, moderate effect: 0.30, large effect: 0.50

SOURCE: U.S. Department of Education, Institute of Education Sciences, National Center for Education Statistics, High School Longitudinal Study of 2009 (HSLS:09), Base Year, First Year Follow-Up, Second Year Follow-Up

*p < 0.05. **p < 0.01. ***p < 0.001