Another examination of table 3 reveals that being a member of the group Laotian/Cambodian does indeed predict a higher GPA and the effect is substantial: Laotian/Cambodian-origin students, after controlling for SES and other factors, earn a half-grade higher GPA. Being Laotian/Cambodian, however, does not increase scores on the standardized test of math, and results in 12 percentiles less on reading. Their success, in other words, is limited to good grades. Walker-Moffett (1995) interprets this result as showing that the Hmong try very hard, and please the teacher, but do not accomplish as much as their high grades indicate.

Additional evidence for this interpretation comes from Lee (1995), in her ethnographic study of Asian-American high school students:

"At Academic High I found that Asian American students were rewarded for being quiet, polite, and respectful. On several occasions I learned from teachers that they had given passing grades to Asian American students who had not earned them. In each case, the student who was passed was described as a quiet and polite student ... similarly, in her study on Hmong students, Goldstein (1985) [unpublished dissertation] found that Hmong students were promoted to the next grade based on their behavior rather than on their academic performance" (pp. 62-63).

I interpret this result as showing that for less privileged children, hard work will take you only so far: It may result in higher grades, but the effect may be limited to higher grades. The effects of poverty and having less-educated parents is very strong. As discussed below there is a lot that can be done for these children, but simply telling them to try hard in school and get good grades is not the complete solution.

Discussion

My conclusions are consistent with other results showing the powerful effect of low SES. Krashen (2002) concluded that despite claims made by Ed Trust, a re-analysis of their data for California showed that high poverty schools that were also high performing, as defined by performance on standardized tests, were nearly non-existent.

My conclusions are also consistent with research showing that high performance on standardized tests of reading is consistently associated with access to print (Krashen, 1993, McQuillan, 1998), and that children of poverty have far less access to print (Neuman and Celano, 2000). This relationship holds because access to print results in more recreational reading, and more recreational reading results in more literacy development (Krashen, 1993). Massive amounts of traditional homework cannot build literacy nearly as effectively as hours of pleasure reading. The obvious part of the cure for children of poverty is more access to print, via better school and public libraries.

Higher SES is also associated with better education in the country of origin. For students who arrived in the US during school age, children of poverty rarely have had quality education in the primary language, a factor that has a strong effect on performance on school in the US. In fact, some middle class immigrant children not only have a strong background when they arrive, but are also provided with tutoring (Krashen, 1996). Bilingual programs are designed to provide this background rapidly for those with limited English proficiency.

5