Record: Oz Guterman and Ari Neuman, “Reading at Home: Comparison of Reading Ability Among Homeschooled and Traditionally Schooled Children” in Reading Psychology 40(2019): 169-190. [Abstract Here].
Summary: Oz Guterman is the senior lecturer of the Department of Human Resources at Western Galilee College where Ari Neuman is also the senior lecturer of education. This is the latest of many articles the pair have been publishing over the last few years. This new one the first to draw on a new set of data they have collected that allows for comparisons between home and conventionally schooled children in Israel.
The authors begin with a strong lit review of the research on homeschoolers and reading, identifying three limitations of the previous work: 1. most of it is conducted retrospectively on young adults, not on children themselves as they learn to read, 2. none of the research examines homeschoolers’ level of general knowledge and intelligence, two things much research has found to be correlated to reading ability, and 3. most of the previous studies have lacked control groups to allow for true scientific comparisons. For this study Guterman and Neuman studied children actually learning to read, age six to twelve, divided into three age groups. They tested these children not only for reading ability but also for general knowledge and intelligence, and also for listening comprehension. Finally, they used a demographically matched sample of children in the school system for comparison, following the method pioneered by Martin-Chang, Gould, and Meuse in 2011. All of this was done in Israel, studying Hebrew-speaking children.
The sample included 101 children, 65 of whom were homeschooled and 36 of whom attended school. 42 were girls; 59 were boys. Breaking the children into age groups and performing a t-test produced no difference between the groups, so age was not a factor affecting any of the results. Neither was gender. Parent education level and family income was the same for both the homeschooled and schooled group.
Both homeschooling and schooling families were recruited directly by the researchers, who enjoyed very high response rates from the families they contacted. Research assistants were trained to develop trusting relationships with families and to administer questionnaires both to parents and to children and to hold an open-ended discussion with both together. The questionnaires included ELUL tests for detection of learning disabilities and Wechsler tests for cognitive ability and general knowledge. The Elul protocol includes tests for reading comprehension, phonological awareness, orthographic awareness, and listening comprehension. The Wechsler test is very common in Israel and has very high reliability coefficients.
Guterman and Neuman performed statistical analysis on the test results, and the findings were as follows. Children who attended school had much higher phonological awareness and higher reading comprehension than homeschooled children. This was especially the case among younger children, though by grade 5 and 6 the gap had narrowed considerably. Difference in orthography and listening comprehension were insignificant. On the other hand, the Wechsler test revealed that homeschooled children had greater general knowledge than schooled children. Interestingly, regression analysis revealed that among homeschoolers, the better readers also had more general knowledge. For schooled children, however, there wasn’t a correlation between the two.
What explains these findings? Ruling out other explanations, Guterman and Neuman speculate that the differences in reading comprehension and especially phonetic awareness likely stem from the different methods used to teach reading. Many homeschooling parents are likely teaching their children to read more “naturally,” rather than focusing on phonics as professionally-trained teachers are taught to do. Furthermore, whereas in school the purpose of reading instruction is to learn to read better, in the home reading might be more instrumental—a means of acquiring knowledge. Homeschoolers don’t read as well, but when they do read they’re reading to learn things, not just to become better readers. The authors freely admit that these are just speculations, and they suggest that future research should look more closely at how exactly homeschooling parents do teach reading and how reading is used in the home.
Guterman and Neuman end by noting several limitations of their research, most notably that it is being performed in Israel among Hebrew-speaking families. It is unclear the degree to which these findings could be generalized to other places, which might not have such a uniform method of reading instruction in the schools and such a small homeschooling population (only about 360 families in the entire country homeschool (p. 170)). They also recognize that they did not disaggregate their homeschool sample by type of homeschooling methodology the way Martin-Chang, Gould, and Meuse did (finding, by the way, that homeschooled children whose mothers provided a more structured curriculum performed a lot better than did `unschoolers.’) and the way they themselves did in a previous study.
Appraisal: This is a wonderful study to have. For years I have been holding up Martin-Chang, Gould, and Meuse as a model for where the homeschooling research into academic achievement should go. This is the first study since 2011 to actually do so. Performing the same assessment on two demographically equivalent samples, one schooled and the other homeschooled, is a great way to avoid the deeply flawed approach of comparing homeschoolers to national averages. As Guterman and Neuman note, their design here is not very complicated. It could and should be replicated in many other locales. If it were, we’d have a much better understanding of homeschooling and academic achievement.
The actual finding, furthermore, is quite surprising and provocative. Heretofore most research has found that, to the degree that there are differences at all in academic achievement between schooled and homeschooled children, the homeschoolers tend to have a slight advantage in reading and a slight disadvantage in math. This study, one of the best ever conducted on the question methodologically, undermines that historic finding, especially for the younger years. However, its finding that the gap closes over time, along with its finding that homeschoolers, though they can’t decode quite as well nevertheless seem to know more than their schooled counterparts, suggests that if there is a reading deficit it does not lead to significant long-term negative consequences for homeschoolers. Guterman and Neuman in their discussion section make more of their finding than is warranted in my opinion. At one point they say that having parents untrained in formal pedagogy teach reading “raises the question as to whether there might be many cases in which such parents are unfit to teach and of the potential damage of nonprofessional homeschooling.” (p. 185) That’s a lot to claim from data that shows that homeschoolers end up, even if they’re not quite as good at reading, scoring higher on intelligence and knowledge measures than their demographic peers in schools.
Milton Gaither, Messiah College