Forcing someone to do something against her will is not only morally reprehensible but more importantly it is illegal. Lately, the news media has been flooded with stories of bride kidnapping in Kyrgyz Republic. Moreover, these stories report that bride kidnapping is so prevalent that some even cite specific numbers based on 'scientific' research. Namely, all of these reports cite studies conducted by Professor Kleinbach (mostly coauthored with fellow sociologists or students), which suggest that over 50% of all marriages in Kyrgyz Republic are non-consensual. For someone who doesn't understand statistics, or for that matter scientific methodology, these numbers either make sense or do not raise any suspicion. For others who make a living by using scientific methodology, these numbers are meaningless. The research conclusions, in particular, are non sequitur.
Professor Kleibach, I am sure, is a great scholar in his field. In fact, his analysis of the causes and consequences of bride kidnapping are probably exemplary. That is not what I am concerned about. I have two problems with his research. One, he either intentionally or unintentionally (as a result of his poor scientific training) makes false inferences about the frequency of bride kidnapping to Kyrgyz population. Second, his claim that educational material he introduced led to a reduction in the prevalence of bride kidnapping is careless and deserves proper criticism.
Where did Kleinbach get his numbers? He got them by going out to certain villages (in one of his most cited papers, it was only one village), which he chose based on previously established relationships (where people have had problems and were willing to talk about bride kidnapping), and conducting surveys of households who were willing to tell their stories. Based on these survey responses, he calculated the frequency of non-consensual marriages. Those are the numbers that we consequently see in the newspapers and other media. What is the problem? The problem is that numbers provided by Kleinbach et al. are true for the sample that he investigated, but they are NOT true for the entire Kyrgyz population. It is a classical example of how one can lie with statistics. We will get to the reasons why Kleinbach may want to lie.
Suppose, you are interested in finding out how many people want to keep the current flag of Kyrgyz Republic. Let's assume Mr. X hates the current flag. Suppose, residents in 90% of villages and towns do not want to change the flag. Only in some towns and villages (town A, B, and C) people wish to change the flag. If Mr. X goes out and surveys people in towns A, B, and C, he will get a result that would support his agenda. If he finds that in the sample (responses from towns A, B, and C) about 80% of people want to change the flag, he can announce to the media that the overwhelming majority wants the current flag gone. Similarly, Kleinbach's claim that incidence of bride kidnapping is high in Kyrgyz Republic, based on his surveys, is false.
When I made a fuss about the reported frequency of bride kidnapping in KG, I've heard people say, "If you are so smart, why don't you tell us your number?" I don't have a number, because to generate that 'number' one needs a carefully designed and properly conducted survey, which requires resources that I do not have. Without getting too technical, one needs to understand that statistics (averages, percentiles, variances, etc..) derived from a sample are meaningful if the sample preserves population characteristics. The best and simplest case of sample selection is a method called 'simple random selection'. This method requires that ever person in Kyrgyz Republic has an equal chance of being selected to the sample. In Kleinbach's studies, none of his surveys pass the minimum requirements of a proper sample selection for population inference.
So, what do we know about bride-kidnapping in KG? We know it is bad for those who were kidnapped. But we do not know how often this happens in Kyrgyz Republic. Kleinbach believes otherwise. As if his initial false claims were not enough, he went ahead and conducted similar surveys in subsequent years. Then, he calculated the incidence of bride-kidnapping in those samples. Using these numbers, he made an outrageous claim that bride kidnapping is on the decline, and his educational efforts have something to do with that. This is not only unscientific, but it is simply insulting to anyone who understands research methodology.
Why would he lie about these numbers? One reason is he lacks statistical training. Another reason is money and fame. This guy gets funded year after year. Moreover, the size of his funds is increasing as he makes more outrageous claims. You may ask, "What's your problem?" My problem is that researchers like Kleinbach may re-divert already tight resources in Kyrgyz Republic from problems that we know are afflicting hundreds of thousands of children (orphanages, immunization, education, prevention of child abuse, and many more) to a problem that may not be affecting a large number of people, but have made policy makers to believe that it does. Every time Kleinbach gets a funding boost from an external organizations that are willing to help Kyrgyz Republic with pressing social problems, another problem of no-lesser importance (or of greater importance) gets underfunded. I am not against studying the problem of bride kindapping, but I am against studies based on False Science.
P.S. Here are the links to Kleinbach studies:
http://faculty.philau.edu/kleinbachr/2004_study.htm
http://faculty.philau.edu/kleinbachr/third_paper.htm
