The Hard Work Hypothesis: Is Doing Your Homework Enough to Overcome the Effects of Poverty?

Yet the Hmong do well in school, and it is undeniable that they are very hard workers. Table 1, from Portes and Rumbaut, reveals, in fact, that the Hmong are the champion homework students of their entire sample, with nearly half reporting two or more hours of homework per day.

**Table 1: Background and Homework**
	school engagement	homework
Mexico	52%	14%
Nicaragua	57%	21%
Vietnam	59%	45%
Hmong	62%	48%
Chinese/other	61%	38%

All data from Portes and Rumbaut, tables 4,1, 8.4
School engagement: percent who feel grades are very important
Homework: percent who report doing two or more hours per day of homework

Their homework time appears to pay off in terms of better grades: Rumbaut (1997) reported that from the sample of children of immigrants in US high schools studied by Portes and Rumbaut (2001), more time spent in homework resulted in better grades (table 2).

**Table 2: Homework and GPA**
Hours	GPA: 1992	GPA: 1996
less than 1	2.3	2.3
1 hr to 2 hrs	2.6	2.6
2 to 3 hrs	2.9	2.8
4 or more	3	3.1

From: Rumbaut (1997)

More Evidence for the Hard Work Hypothesis: Background Counts When SES is Controlled

Table 3 presents a multiple regression analysis from Portes and Rumbaut (2001), examining predictors of grades and standardized test results for children of immigrants in high school.

Multiple regression is a very useful statistical tool that allows us to examine the impact of different predictors, holding the others constant. For example, table 3 tells us that high school students who have higher socio-economic status (SES) score higher on a test of reading. The "regression coefficient" for SES is 6.88. SES was measured on a five-point scale (-2 to +2); for each point higher in SES, students scored 6.88 higher on the reading test. Thus, students from the highest SES group scored about 34 points higher in reading than students from the lowest group.. The regression coefficient represents the impact of SES without influence of other factors, that is, when interpreting the impact of SES, we can pretend students were identical in all other ways.

In table 3, SES is a strong predictor of all three measures, reading, math and grades. Note, however, even after controlling for SES, as well as for other predictors, such as the SES of one's friends, background is still a significant predictor. Being Mexican, for example, predicts lower test scores (15 percentiles lower in reading) as well as a quarter of a grade lower grade point average. Being Chinese/Korean predicts 3/4 of a grade higher GPA, 13 percentiles higher in reading, and a spectacular 23 percentiles higher in math.