3
| Predictor | READING | MATH | GRADES |
| Age (range 12-18) | -3.24*** | -4.10*** | -.10*** |
| Sex (1 = female) | 2.63* | 1.62* | .33*** |
| Region (1 = CA, 0 = FL) | 3.85* | -8.46** | 0.27 |
| Parental SES (range -2 to +2) | 6.88** | 5.8*** | .18*** |
| Intact Family (1 = both parents) | 0.9 | 3.45** | .18*** |
| US Born | 6.06*** | 0.56 | -.16*** |
| Long Term Resident* | 6.23*** | 0.02 | -.12* |
| Fluent Bilingual (1 = fluent bi.) | 2.3* | 1.93* | .06* |
| Limited Bilingual | -12.14*** | -7.14*** | |
| Parent-Child conflict (1-4) | -2.15** | -3.12*** | -.16*** |
| Second-generation friends (1-3) | 1.88* | 3.84* | -.06** |
| Inner City (1 = inner city, 2 = not inner city) | -2.07* | -6.21** | 0 |
| Average Student SES (free/reduced lunch) | .2*** | .10** | -0.06 |
| Chinese/Korean | 12.97** | 23.22*** | .76*** |
| Columbian | 0.07 | -0.45 | -0.05 |
| Cuban, private school | 7.14* | 5.6 | .25** |
| Cuban, public school | -1.64 | -3.48* | -.14* |
| Filipino | -1.69 | 3.39 | .17* |
| Haitian | -8.29** | -7.18* | 0 |
| Laotian/Cambodian | -12.41** | 2.1 | .47*** |
| Mexican | -14.7*** | -11.52*** | -.25** |
| Nicaraguan | -3.71 | -2.62 | -0.1 |
| Vietnamese | -3.21 | 15.79*** | .50*** |
| West Indian | 1.07 | -4.94 | -0.01 |
| r2 | 0.29 | 0.25 | 0.24 |
* = moderate effect, ** = strong effect, *** = very strong effect
Length of residence in US: 1 = less than 5 years; 2 = 5-9, 3 = 10 or more, 4 = native born.
Range of grades = 0 to 5, mean = 2.52
A look back at table 2 gives a plausible explanation: Chinese/other Asian background students do a lot more homework than Mexican background students, and are more engaged in school. This seems to say that homework counts, that social class is not everything, and suggests that homework can at least make up for some of the disadvantages poor children have.
The Counterevidence: Another look at the Boat People. How well did they really do?
A closer look at research on the Boat People reveals that they did well, but were not spectacular. As noted above, their overall GPA was 3.02, but much of this was due to high performance in math: Without math, their average drops to 2.64 (Caplan et. al., 1989, p. 67).
For the 96 high school students studied, CAT math scores were very high, but their CAT scores for language and reading were a more ordinary 46th percentile. This is very good for students who have only been in the US for three and a half years, but not spectacular. (It should be noted that Caplan et. al. reported the results for CAT reading and language tests combined. They are separate tests, and the reading test is considered more demanding and a better test of academic language.)
3