Dual-language elementaries score LOWER on ELA — but that's not the whole story

Thesis

158 NYC schools tag dual-language programs. They average 51.1% grade-5 ELA proficiency vs 60.4% at other schools — a 9-point deficit. But these schools serve disproportionately more ELL students. Once you control for student population, dual-language may be neutral or positive.

Supporting findings

• Dual-language tagged ES (n=158): avg grade-5 ELA proficient 51.1%.
• Other ES (n=623): 60.4%. Gap: -9.3 points.
• But: dual-language schools serve disproportionately more ELL kids (the program is for them).
• Need a proclivity-decile-adjusted comparison: dual-language vs non-dual-language WITHIN each decile.
• Some dual-language schools are stellar (PS 1 Bergen, PS 84 Lillian Weber). Others struggle. Is the variance higher in dual-language?
• Look at the ELL-vs-NEVER_ELL gap WITHIN dual-language schools — is it smaller there (the whole point of dual-language)?
• Compare with research literature: Lindholm-Leary at SJSU has 20+ years of dual-language outcome studies. NYC schools should match her findings.
• Is the dual-language label being applied to schools that aren't really dual-language (just bilingual / ESL with marketing)? Verify via a sample of school sites.
• What languages dominate? Spanish-English is most common but Mandarin, Haitian Creole, Bengali, French exist.
• Is the math gap (if any) smaller than the ELA gap? Math is more language-neutral and should be a fairer comparison.

Reporting directions

• Build a regression: ela_grade5_prof ~ dual_language + proclivity_score + pct_ell + borough.
• Interview the principal of PS 1 Bergen (highest-performing dual-language ES).
• Compare 5th grade ELA at dual-language schools who ALSO test in Spanish (some do) — measure proficiency in BOTH languages.
• Interview Kathryn Lindholm-Leary on what to expect from NYC's program portfolio.