When Homeschoolers Turn Violent: Conclusion

Screen Shot 2014-03-02 at 5.36.11 PM

HA note: The series conclusion of “When Homeschoolers Turn Violent” is written by Rachel Coleman, Rachel Lazerus, and Dr. Chelsea McCracken from Homeschooling’s Invisible Children.


Homeschooling communities assume that homeschooled kids escape the pernicious influences of the wider world — if violent behavior is something that you catch, like a disease, homeschooling families assume that they can protect their children from exposure. Some homeschooling advocates also claim that those who have been homeschooled properly will never turn out to be violent criminals or have mental health issues.

The assumption that homeschoolers are essentially good kids is also held by members of the general public. Joshua Komisarjevsky was sentenced to only nine years for his role in burglarizing 18 homes, despite the maximum statute being ten years for each offense. When he came before the parole board, he was seen as “young, white, bright, home-schooled, remorseful, never identified as a person with high mental health needs.”[1] He was paroled in April 2007, and invaded the Petit family home on July 23rd, 2013.

Homeschool advocate Brian Ray implicitly endorses this view, writing:

Discussion with leaders in the homeschool movement of the past several years of news and careful Internet searches reveal 2 and possibly 3 homeschooled teens (ages 14 to 17) who were accused or convicted of homicide during the past 8 years. …If home-based education of teens of ages 14 to 17 comprises 3% of that age-group population and calculations are adjusted for number of homicide offenders and number of years (with a liberal estimate of 3 during 8 years), then homicide offenders of homeschool students ages 14 to 17 would be about 0.004 per 100,000. If true, this would mean that the general-population teen of this age is 2,500 times more likely to commit homicide than a home-educated teen.

In this conclusion post to our series on homeschoolers who have turned to violence, we would like to explain why it is both unfair and unscholarly to compare homicide rates in the way Ray does — and how these flawed assumptions make homeschooling families and communities less safe.

First, we would like to note that it is unscientific to conclude from anecdotes that there were only two or three homeschool homicides during the previous eight years. For one thing, our list includes more than twice this number, and we do not assume it is a complete list. Not every youth homicide makes the news, not every news story on a youth homicide discusses the offender’s educational background, and not every story that makes the news is easily accessible on the internet. To accurately compare the rate of homeschool homicides to the national rate, one would need to access the criminal records for all youth homicides within a given time period and look at the educational background of each. It is irresponsible to speculate on youth homicide rates without this data.

Even more importantly, comparing youth homicide rates between the homeschool population and the general population would not actually tell us anything about whether homeschooling itself affects the youth homicide rate. When comparing homicide rates between two populations, it is crucial to consider all of the ways the populations differ. In other words, if the homicide rate is indeed lower among homeschooled youth, which it may be, this might be because homeschooled youth differ from youth in the general population in various ways that are unrelated to school choice.

When comparing crime rates in two populations, it is important to take into account various factors that might impact the crime rate. For example, one cannot compare Alabama’s high crime rate with New Hampshire’s low crime rate without taking into account the difference in poverty rates: Alabama has one of the highest poverty rates in the country while New Hampshire has one of the lowest. According to the National Center for Child Death Review, “Major contributing factors [to youth homicide] in addition to poverty include easy access to handguns, involvement in drug and gang activity, family disruption and school failure.” Other factors which may also contribute to crime rates include education level, community involvement, availability of mental health services, condition of law enforcement, etc. When comparing two populations, all of these factors must be taken into account.

The homeschool population varies from the general population in a number of ways. The homeschool population tends to be more rural than urban, and homeschooled children are more likely to live in households with two parents than are children in the general public. It has traditionally been thought that parents who homeschool tend to be wealthier and better educated than average (although this has recently been called into question). All of these factors correlate with lower crime and homicide rates. Without correcting for these various factors, it is impossible to know whether homeschooling might play any role in lowering homicide rates.

Many homeschool families also have high levels of parental involvement, which is not easily quantifiable. Parental involvement leads to higher educational attainment in any educational setting and also correlates with lower crime rates and lower engagement in risky behaviors. As a result, lower homicide rates among homeschooled youth could be a result of high parental involvement rather than of homeschooling per se — and highly involved parents would affect their children’s lives in this way regardless of the educational method they chose.

Could these background factors be corrected for? Yes. To do so, one would first have to collect accurate data on homeschooled students and their families. Then, one could compare the homicide rate among homeschooled youth with the homicide rate among youth in the general population with the same background factors. In other words, if homeschooled youth are more likely to be rural, more likely to have two parents, more likely to live in families with slightly higher incomes and educational attainment, and more likely to have high parental involvement, the homicide rate among homeschooled youth would need to be compared to the homicide rate among a sample of the general population that is also more likely to be rural, more likely to have two parents, more likely to live in families with slightly higher incomes and educational attainment, and more likely to have high parental involvement. That way one could isolate any effect educational method might have.

As the director of the National Home Education Research Institute, Ray has conducted a number of studies on homeschoolers; however, he has consistently failed to correct for these background factors. In several of his studies, for example, he gathers testing data from a nonrepresentative sample of highly educated, high income, highly involved homeschool families and then deceptively attributes their predictably high test scores to the fact that they were homeschooled. (For more, see reviews of Ray’s work here.)

When a youth homicide occurs, the student’s educational background, whether public, private, or homeschool, is not generally the causal factor. Attending public school does not cause an adolescent to commit a homicide any more than being homeschooled causes an adolescent to commit a homicide. Rather than focusing on which educational method is correlated with the lowest homicide rate, we should instead study how various factors may contribute to adolescent homicides in any educational setting. We prefer to focus on keeping youth safe, not on making an ideological point.

We do not have the data to know whether homicide rates are higher or lower among homeschooled youth, or what role homeschooling plays. However, there are some factors we may observe as common themes in homeschool homicides. In many of the cases we have collected, homeschoolers who commit homicide come from families with extreme religious or ideological beliefs (patriarchy, white supremacy, anti-government views, etc.). Many of these youths have easy access to firearms in their homes. They may have untreated mental health issues or developmental disabilities. Several youths come from disrupted homes, either through adoption or divorce. Parental abuse and neglect, as well as drug and alcohol abuse, also play a role. Though in some cases these factors may correlate with homeschooling, this does not imply that homeschooling causes homicides — the same factors have been implicated in homicides committed by youths who attend school.

It is a common and understandable impulse to try to keep ourselves and our community safe by believing that violence could never happen to us. Unfortunately, this reaction is not productive. The best way to keep homeschoolers safe is not to deny that there could ever be any problems, but rather to learn about the factors that contribute to problems and to be on the lookout for mental illness and abuse among homeschooled students, homeschool grads, and homeschool parents.

[1] Rep. Mark Lawlor, Chairman, CT Judicial Committee in The Cheshire Murders, HBO documentary, 2013.

View the case index here.