Post a new comment on this article
Post Your Discussion Comment
Please follow our guidelines for comments and review our competing interests policy. Comments that do not conform to our guidelines will be promptly removed and the user account disabled. The following must be avoided:
- Remarks that could be interpreted as allegations of misconduct
- Unsupported assertions or statements
- Inflammatory or insulting language
Why should this posting be reviewed?
See also Guidelines for Comments and Corrections.
Thank you for taking the time to flag this posting; we review flagged postings on a regular basis.close
Is Cause and Effect Conclusion Justified?
Posted by Todd_Ryan on 21 Feb 2012 at 16:44 GMT
The authors attempted to determine the effectiveness of abstinence education by comparing teen pregnancy rates, birth rates, and abortion rates for U.S. states for the year 2005 as a function of the degree to which each state emphasizes abstinence in its sex education programs. They do this by categorizing each state’s emphasis on abstinence education into four categories with zero being the least emphasis and three being the most emphasis.
Their statistical analysis found that teen pregnancy is positively correlated with abstinence education, and they conclude from this correlation that abstinence education is ineffective at reducing teen pregnancy and may increase it.
However, I am concerned that the authors draw too strong a conclusion of cause and effect from this correlation.
The federal government began promoting abstinence education in earnest in 1996 with the passage of Title V, Section 510 of the Personal Responsibility and Work Opportunity Reconciliation Act of 1996. Thus using 2005 data allowed sufficient time for abstinence education to yield observable effects. However the authors did not report a similar analysis for the available data prior to 1996.
I examined previous years using the authors’ laws and policy level categories from their Table 2 and teen pregnancy, birthrate, and abortion data from the same Guttmacher Institute report cited by the authors. State assignment to the categories would obviously change as curricula change over time, but keeping the categories the same allows a look at where each category was before Title V initiatives (assuming abstinence education was not taught in those states prior to 1996, which may not be true).
After reproducing the 2005 trends shown in Figures 1 and 2 I examined the years 1996 and 1988 and found that category 3 states have always had higher teen pregnancy and birth rates than category 0, 1, and 2 states. Thus it is not clear that Title V changes in sex education curricula are the cause of higher pregnancy rates for category 3 states.
I also plotted the change in teen pregnancy rates for the time spans of 1988 to 1996 (prior to Title V) and 1996 to 2005 (after Title V). Prior to Title V category 1 states had a significant decrease in teen pregnancy relative to all other states. After Title V category 1 states continued to show the largest change in teen pregnancy rate, but that change is not significantly different from change in category 3 states.
When data prior to Title V passage is considered it suggests that the correlation between abstinence education and teen pregnancy rates reported by the authors is not caused by changes in sex education curricula but something else.
Any such analysis has limitations, but in my opinion the
authors’ strong claim to cause and effect is not justified unless they can determine the effect of curriculum changes relative to a starting point prior to the change.
In our paper, we were careful to state that we have shown correlation, not causation — I.e. we have not shown that abstinence emphasis is the cause of increased teen pregnancy, but that it is significantly positively correlated with it.
However, we also note that causation would result in a correlation. If abstinence education actually caused abstinence behavior then, after accounting for other factors (outlined in the paper), more emphasis on abstinence should be correlated with lower teen pregnancy rates. The actual (significant) correlation is in the opposite direction.
We did not analyze data prior to Title V emphasis because we do not know what states were doing prior to that time with respect to sex education. Mr. Ryan suggests that level 3 states have always had high teen pregnancy and so abstinence education did not "cause" these rates, instead they were already high. The problem with such an analysis is that we do not know how much states emphasized abstinence education in the 1990s and before, so whether teen pregnancy rates were high prior to Title V gives no information on whether abstinence education is “causally” correlated as suggested.
From an experimental standpoint, examining the effects of abstinence education before and after institution would be ideal. However, institution of title V is not equivalent to institution of abstinence education, and making this assumption is problematic. The proposal to look at the difference in pregnancy rates before and after institution of Title V might get at the effect of funding but not at the effect of abstinence education per se. To test for the effect of education specifically the analysis would have to compare state data before and after a change in their respective state laws and policies.
In any such analysis, nationwide trends in teen pregnancy rates and the influence of other factors would have to be statistically accounted for.
The primary focus of our study was to examine the most recent data on teen pregnancy for a correlation between the sex education approach prescribed by individual states and the learning outcomes of students (as measured by teen pregnancy rates) in these states.
As outlined in the paper, the correlation data unequivocally indicate that abstinence-only education is not successful. Instead, it is associated with higher teen pregnancy rates than comprehensive sex education, even after accounting for other contributing factors.
Kathrin Stanger-Hall and David Hall
RE: RE: Is Cause and Effect Conclusion Justified?
Professors Stanger-Hall and Hall concluded in their paper that, “these data show clearly that abstinence-only education as a state policy is ineffective in preventing teenage pregnancy.” They conclude this because teen pregnancy rates are positively correlated with emphasis on abstinence. They reason that if abstinence is effective then states emphasizing abstinence should have a lower rate of teen pregnancy, but they observe the opposite.
The problem with this conclusion is that they have no baseline from which to compare. As they state, “We did not analyze data prior to Title V emphasis because we do not know what states were doing prior to that time with respect to sex education.”
Teen pregnancy rates have been decreasing in all states since about 1990 including states that now emphasize abstinence. The decrease in teen pregnancy rates in states now emphasizing abstinence has been similar as in states that now do not emphasize abstinence. The rate of change has been about the same in both groups of states regardless of whether they emphasize abstinence or not. This data is available from the same sources used by the authors.
However, the states that now emphasize abstinence have tended to have higher teen pregnancy rates than other states since 1988 (as far back as the dataset goes), which is before the national decline in teen pregnancy rate began. So today, the states emphasizing abstinence still have higher teen pregnancy rates (on average) despite having a similar rate of decrease in teen pregnancies as other states. They started out with higher rates before the national decline in teen pregnancy rate.
The only way that states now emphasizing abstinence (as a group) could have a lower teen pregnancy rate today than other states would be if their rate of decrease in teen pregnancies were greater than other states.
Thus, the positive correlation that the author’s observe today says nothing conclusive about sex education approaches.
In their response to my first comment above the author’s state, “the correlation data unequivocally indicate that abstinence-only education is not successful.”
The fact that states emphasizing abstinence and states not emphasizing abstinence show similar rates of decline in teen pregnancies since 1990 does not support that conclusion. The similar rates of decline could suggest that both approaches to sex education are equally effective. However, as the authors state, we don’t know how those states taught sex education in prior years. Thus these trends or the correlation observed by the authors reveal nothing conclusive or even suggestive about sex education approaches.
As the authors state in the response above, to see the effect of abstinence education you need to examine teen pregnancy rates in individual states before and after institution of the program.