In a recent publication, Ying Chen and other researchers[1] reported that homeschoolers were more likely to become good citizens. At the same time, they also determined that homeschoolers were less likely to attain a college degree by the time they reached their mid-20s and might be more inclined to develop post-traumatic stress disorder (PTSD) as young adults. While the study generally provided some good news about homeschooling’s effectiveness, it would need to be replicated to determine whether its conclusions really reflect the characteristics of the homeschool population. This is due to limitations in the study’s participants and research design. Additionally, the PTSD finding cannot be relied upon due to the measurement tool used.

Subjects and Method

The Chen study[2] included a cohort of over 12,000 students (each student had at least one parent who was a nurse). Roughly 1,300 of the participants were educated in private schools, 1,100 in religious schools, 200 in homeschools, and 9,500 in public schools. The cohort included students who a) experienced a high standard of living, b) were mostly non-Hispanic Whites, and d) were physically healthy. It also included more females than males. Private and religious school students generally experienced the highest standard of living growing up. However, those who attended religious schools or homeschools were more likely to live with both biological parents, have frequent family dinners, and engage less often in smoking and drug use. Over the course of the study, the average age of the participants ranged from 14½ to 25.

Between 1999 and 2010, the researchers repeatedly collected data by administering the Growing Up Today Survey to each participant on an annual basis. The survey inquired about their psychological well-being, social engagement, character strengths, mental health, health behaviors, and physical health. Additionally, on one occasion in the middle of the study, the authors used a screening test for post-traumatic stress disorder PTSD called Breslau’s 7-Item Short Screening Scale.


In their analysis of the data, the authors employed primary methods that allowed them to evaluate the differences between groups and then followed up with secondary or “corrective” analytic procedures to help ensure the accuracy of the initial findings. The impact of using this approach was that some of the findings that were found to be significant during the first analysis were not found to be significant during the second one. Findings needed to pass both levels of analysis to ensure real differences between groups. For example, when comparing the private and public school groups, the first level of analysis indicated that the private school students reported higher levels of forgiveness than the public school group. However, because the second stage or analysis did not validate the first level of analysis, it was not possible for the authors to verify whether the private school students in the study were actually more forgiving.

The same approach was used when comparing students from religious and public schools. During the first level of analysis, it appeared that the religious students were more likely to attend religious services, become registered voters, engage in binge drinking, and have fewer sexual partners. Additionally, religious students were less likely to experience obesity. However, because none of these differences could be confirmed during the second analysis, it was uncertain whether any real differences existed between the two groups.

In contrast, the comparisons involving homeschool and public school students yielded findings that withstood both levels of analysis. For example, the first level of analysis indicated that the homeschool group seemed more likely than the public school group to attend religious services, volunteer frequently, forgive, have fewer sexual partners, have a sense of mission in life, and experience PTSD during young adulthood. They were also less likely to use marijuana and attain a college degree by early adulthood. However, the second analysis could only verify that the homeschool students were more likely than the public school group to attend religious services, volunteer, and forgive while being less likely than the public school students to attain a college degree by the time they reached their early to mid-20s.


One issue that makes it difficult to know whether these findings apply to homeschoolers in general is the fact that the homogeneous sample of 12,000 individuals (i.e., mostly affluent non-Hispanic White persons) was not representative of the more diverse US school-age population (White, 46%; Hispanic, 28%; Black, 15%; Asian, 5%; two or more races, 5%; American Indian/Alaska Native, less than 1%; Pacific Islander, less than 1%),[3] especially when considering that every student in the study had a parent who was a nurse. Regarding homeschoolers, the study’s results might perhaps have generalized satisfactorily to the homeschool community when it involved mostly non-Hispanic White families during the 1980s and 1990s. However, because the racial and income diversity of the homeschool population now more closely reflects the nation’s student population in general,[4] the findings in the Chen[5] study are perhaps useful for understanding only a fraction of today’s homeschoolers. Furthermore, this same criticism would, to a degree, apply to the other groups in the study because they, too, included mostly affluent, non-Hispanic White participants. Consequently, it is difficult to gauge the overall contribution that this study makes to understanding the broader picture of how these different types of schools prepare students to function as young adults.

Another limitation of the study surrounds the authors’ implication that, during young adulthood, the homeschooled students in the study may have experienced a higher likelihood of PTSD based on the results of Breslau’s 7-Item Short Screening Scale. [6] Even though this finding was found to be significant only in the first level of analysis and not the second, it seems important to clarify that, even if this finding had held up in the second analysis, its meaning would have been highly suspect because the Breslau is an obsolete screening test.

Screening tests are typically developed to help determine whether more comprehensive diagnostic testing is warranted—not to make diagnoses.[7] A screening test, for example, may include 10 questions that are taken directly from a 50-item diagnostic test. The shorter list of questions that comprise the screening test is then validated by determining whether it generally identifies the same individual traits that the diagnostic test identifies. If it is determined that the two tests identify the same traits in a client, the screening test is said to have adequate validity. However, the brevity of screening tests prevents them from being as consistently accurate as the longer diagnostic ones. As such, screening tests more often identify problems where none exist or fail to identify problems that do. That is why, in clinical practice, screening tests are used to signal the need for longer diagnostic tests to be administered—not to diagnose problems.[8] Even so, in research studies, scientists often use screening tests instead of lengthier diagnostic ones because they are more rapidly administered and scored. When considering this, it makes sense that the researchers in the current study used the Breslau screening test, because they needed to evaluate more than 12,000 individuals for possible PTSD. At the same time, it is misleading to imply that participants in the study actually had PTSD based on their Breslau results.

Additionally, three of the seven items on the Breslau are now obsolete as a result of the diagnostic criteria for PTSD being changed in 2013.[9], [10] But the researchers reported that each item on the screening test—including the three obsolete ones—represented a PTSD symptom and that participants probably had PTSD if they indicated that they experienced six of those seven symptoms. Consequently, it seems that this particular finding offers little, if any, understanding of the prevalence of homeschooled individuals with PTSD, even from a screening perspective.


Because the sample of participants in the Chen study was made up of affluent non-Hispanic White individuals who had parents that were nurses, it is difficult to know how widely the results could be applied to the overall homeschool population. Even so, the study yielded some interesting information about the homeschoolers who did participate. For example, it demonstrated that homeschooled individuals experienced psychological well-being in young adulthood.

Homeschoolers also seemed less likely to obtain a college degree. Possible explanations for this finding are that, at the time when data were being gathered for the Chen study between 1999 and 2010, homeschool students often faced additional entry requirements from colleges and universities[11] or financial aid restrictions[12] that students from other schools did not encounter. Consequently, it should not be surprising to learn that fewer homeschooled individuals attained college degrees two decades ago at a time when it was more difficult for them to gain entry to institutions of higher learning in the first place. In contrast, almost two and a half decades after the first data in the Chen study were collected, college admissions officers now search for ways to attract homeschool students because they are often academically advanced, curious, and diverse.[13] Furthermore, because parents often choose to homeschool children with disabilities, it is reasonable to assume that those disabilities sometimes slow or prevent the students from attaining college degrees.[14]

Regarding the PTSD finding, it should essentially be ignored because the researchers used a deficient and obsolete test to identify whether the participants had PTSD.

Past studies demonstrated that parents could do a good job of preparing their children academically.[15] Now, Chen’s more recent longitudinal study shows that homeschool parents can be effective at preparing their children to be contributing citizens, as well. However, further follow-up study is needed to more fully understand homeschooling’s impact on the development of young people.