Homeschool Parents and Satisfaction with Special Education Services

ABSTRACT

Homeschooling is controversial for a variety of reasons. One concern is whether families are sufficiently equipped to serve students with disabilities. We investigate this issue by assessing parental satisfaction with the special education services that their child is receiving in various educational sectors (e.g., homeschool, traditional public, public charter, and private). Using a nationally representative sample of U.S. households from the National Household Education Survey, we find that parents who homeschool are more satisfied than parents of children in traditional public schools and a variety of private schools with the special education services that they are receiving. Despite obvious selection bias in our sample, we view parental satisfaction as one of many important indicators for the quality of special education services. The results from this study suggest that homeschooling is a potentially beneficial option for serving students with disabilities, though additional research examining other student outcomes would be invaluable.

KEYWORDS:

• homeschooling
• parent satisfaction
• school choice
• special education

Notes

1. Our use of the term homeschooling does not apply to students who are homebound due to their disability.

2. 34 CFR § 300.131 (2004).

3. 34 CFR § 300.137 (2004).

4. We estimate linear regression models for all of our dependent variables. Although the only continuous dependent variable in our models is a measure of overall satisfaction, ordered logit estimation of the other discrete variables yielded similar results as the linear regression models. Hence for ease of interpretation, we report linear regression coefficients.

5. One might worry that the sample sizes for homeschool and private-school families is low, given that they collectively make up only about 5% of the sample. These small sample sizes should not bias our estimates but only make it more difficult to detect statistically significant results. Thus, we can be even more confident that the differences that we have detected across these sectors is material and not due to random chance.