A Lesson in Data and Analysis for the New York Times

Yesterday, The New York Times published a front-page article, "Scant Drop in Abortion Rates if Parents are Told," that reported the newspaper's conclusion that recently-enacted parental involvement laws have not reduced the incidence of abortion among teens.[1] On its surface, the newspaper's statistical analysis appears convincing, but a closer look at its data and methodology call its conclusions into question.

The article'sreporters tracked the percentage of pregnancies among girls under age 18 that end in abortion before and after the passage of parental-involvement legislation in six states. The passage of such legislation, according to the Times' analysis, has little effect on the percentage of pregnant minors who obtain abortions. Furthermore, says the article, following the enactment of parental-involvement legislation, minors' childbearing decisions continued to track those of women ages 18 to 19, who are not directly impacted by parental-involvement legislation.

This Times' analysis, however, contains several significant shortcomings. First, it examines data from only six of the approximately twelve states that have passed parental involvement laws since the mid-1990s. Second, its abortion data come from state health departments, which researchers agree tend to be unreliable; academics who study the incidence of abortion nearly always use data from either the Centers for Disease Control (CDC) or the Alan Guttmacher Institute (AGI) because these sources employ more reliable collecting and reporting mechanisms. Finally, the Times analyzes the percentage of pregnancies that end in abortion rather than the percentage of teens who have abortions. Because relatively few teens give birth each year, the Times' measure can fluctuate dramatically, making the data difficult to analyze properly.

Furthermore, the Times ignores the likelihood that the presence of parental-involvement laws may reduce abortions not only by influencing the decisions of girls who are already pregnant, but also by reducing the number of teenage girls who become pregnant at all. An analysis of the percentage of total teen pregnancies that end in abortion will not lend any great insight into the effectiveness of parental-involvement laws in reducing abortions by changing minors' sexual behavior.

Indeed, a more rigorous analysis of parental-involvement laws, such as I performed last year, tells a considerably different story. Using data on teen abortions from the Centers for Disease Control (now updated to include every year until 2002, the last year for which data are available) and population data from the U.S. Census Bureau, I calculated more accurate teen abortion rates than is possible with the Times' data. Teen abortion rates reflect the approximate likelihood that a girl between the ages of 13 and 17 in a particular state will undergo an abortion in a particular year.

With this data, I examined teen abortion rate in the states that the Times reporters discussed in their article. Because Arizona's parental-consent law took effect in 2003 and CDC data is unavailable for years after 2002, I was unable to include Arizona in this analysis. However, in three of the five other states analyzed by the Times, I found significant reductions in the teen abortion rates after the passage of parental-involvement laws. In Texas, the teen abortion rate fell 25 percent since the passage of a parental-notification law in 2000. Both Virginia and South Dakota passed parental notification laws in 1997, and since that time their teen abortion rates each declined by over 33 percent.

It is true that in the remaining two states the Times examined, Idaho and Tennessee, the passage of parental-involvement laws had little immediate effect on the abortion rate. However, additional information about each state provides some important context, which is ignored in the Times article. Idaho already had one of the lowest teen abortion rates in the country prior to the passage of its parental-consent law. Similarly, while Tennessee's teen abortion rate fluctuated little in the years following the passage of its parental-consent law in 2000, its teen abortion rate fell sharply in the year before the passage of the law. Tennessee's parental-consent law may have played a role in preserving this decline.

Evidence from states other than those included in the Times article also supports the effectiveness of parental-involvement laws. In the early 1990s, Mississippi, Minnesota, and Nebraska passed parental-involvement laws.[2] By the end of the decade, the teen abortion rates in all of these states fell by half. Further research shows that parental-involvement laws reduce the abortion rate among teens but have considerably less impact on the overall abortion rate.[3] When teen abortion rates decline and overall abortion rates remain fairly constant, it is safe to deduce that parental-involvement laws are primarily responsible for the declines in teen abortion rates-and not broad shifts in values or mores that happen to be correlated with the passage of such legislation. Finally, other academic studies published in peer-reviewed journals-also ignored by the Times reporters-provide solid statistical evidence that parental-involvement laws reduce the incidence of abortion among minors when demographic and economic factors are held constant.

It is regrettable that the Times reporters refused to acknowledge academic research that contradicts their conclusions. This continues the newspaper's trend of poor reporting on abortion statistics over the last decade. For example, during the 2004 election season, the Times reported Glen Harold Stassen's erroneous finding that abortions had increased during the George W. Bush's presidency. When the Alan Guttmacher Institute later released more comprehensive data showing that abortions had actually declined since President Bush's inauguration, the Times was among the media outlets that failed to report the finding, much less correct its own record.

The Times reporters, like Stassen, ground their conclusions in data from the health departments of a small sample of states. The Times' analysis also overlooks much contradictory evidence rooted in superior data. Policymakers and citizens should have access to conclusions based on the highest quality data and research methodology. Researchers, policy organizations, and media outlets like The New York Times should be committed to upholding the highest standards for the research that they present to the public. In this case, the "The Paper of Record" fell far short of that ideal. The Times could begin to remedy that failure by making public its reporters' datasets and methodology, which so far it has been unwilling to release.

Michael J. New, Ph.D. is an Assistant Professor at the University of Alabama and a former visiting fellow in The Heritage Foundation's Center for Data Analysis.