100 likes | 276 Views
Study Motivation. Immigrant paradox very well-established: (i.e., Fuligni & Yoshikawa, 2004; Glick & White, 2003; Kao, 1999; Pong & Hao , 2007; Portes & Rumbaut , 2001; Rosenbaum & Rochford , 2008; Suárez -Orozco & Suárez -Orozco, 1995)
E N D
Study Motivation • Immigrant paradox very well-established: (i.e., Fuligni & Yoshikawa, 2004; Glick & White, 2003; Kao, 1999; Pong & Hao, 2007; Portes & Rumbaut, 2001; Rosenbaum & Rochford, 2008; Suárez-Orozco & Suárez-Orozco, 1995) • Focus on adolescence, given larger achievement gap • Yet, some ‘puzzle pieces’ that might contribute to the immigrant paradox have not been examined… • Is the immigrant paradox simply a measurement problem? • How ‘real’ are identified differences across generations? • Does generational status moderate the mediated relations among key inputs to achievement?
CFA Note: model tested for 1st vs. 2nd vs. 3rd gens. separately; Gender is statistically controlled
Data Source • NYCASES: New York City Social and Academic Engagement Study (PI: S. Sirin) • NYCASES surveyed 517 high school students (55.7% female, n = 288; M age = 15.7, S.D. = 1.05) • This study examined 10th graders • Diverse in terms of race/ethnicity: • (European American = 3.1%, n = 16; Black = 19.6%, n = 98; Asian = 16%, n = 80; Latino = 43.3%, n = 217; and Multiracial = 18%, n = 90). • Nearly evenly split by generational status: • 1st generation = 33.1%, n = 169 • 2nd generation = 34.4%, n = 176 • 3rd generation = 29.7%, n = 152 • This split ideal for this study of gen differences! • Achievement measure: transcript-based 1-100 cumulative measure of composite academic achievement, constructed for NYCASES
RQ#1 : ‘Apples to apples’ or ‘apples to oranges’? RQ#1: Unknown if measurement differences explain IP? • Good model fit, loadings & reliability estimates for each measure (Beh SE: Compliance, Beh SE: Effort, Supportive School Relations) • Measurement invariance (‘apples to apples’) established across generations • Therefore, measurement error does not explain the immigrant paradox • Key inputs to achievement are invariant (not biased) across generations & can be measured the same way w/ same meaning • Establishes ‘apples to apples’ comparisons for RQs 2 & 3, as know the measures function equally across groups
RQ#2: Latent mean comparisons RQ#2 : Do 1st vs. 2nd vs. 3rd gen youth have different levels of ‘academic inputs?’ A to RQ2: Yes! Latent means are more precise, particularly after establishing MI
RQ#3: Do the mediated processes to achievement differ by gen? RQ#3: Does gen status moderate mediated relations among inputs to achievement? • Preview of figure in next slide • Generally, students who perceived more school support more likely to exert effort and comply with school rules -> better achievement. • Supportive school relations played a larger role in fostering effort for 1st generation & compliance for 2nd and 3rd generation • Effort engenders achievement more than compliance (for all groups) • SSR differential “downstream’ (mediated) impacts on achievement across generations, in way that reinforces immigrant paradox
* Denotes statistical significance Mediated Relations by Gen. SSR -> Effort -> Achievement (1st β = .24*; 2nd β = .16*; 3rd β = .13*) SSR -> Compliance -> Achievement (3rd β = .08*)
Differential Impacts of Key Inputs • General support for importance of school support for achievement • Key finding: SSR fosters effort for 1st gen vs. compliance for 2nd & 3rd gen • SSR elicits differential responses (across generations) that differentially shape achievement • The same amount of SSR provides more ‘bang for buck’ [in terms of achievement] for 1st gen. than for 2nd and 3rd gen youth • ‘Key inputs’ to achievement more powerful for 1st gen -> contributes to IP • SSR not a ‘silver bullet’ across generations • Policies/practices targeting SSR may yield less achievement gains for 2nd & 3rd gen youth • Unresolved question: If SSR not the critical input, how can we foster effort & achievement among 2nd & 3rd gen youth?
Limitations & Future Directions • Did not control for SES (yet earlier evidence suggests IP is not solely SES) • National as well as racial/ethnic differences not controlled, important given impact of pre- & post-migratory conditions & racial/ethnic differences in the IP (Crosnoe & López-Turley, 2011; Suárez-Orozco et al., 2008) • Cross-sectional analyses of NYCASES data