Research results can often be misleading. Although methods employing statistical measures tend to be more objective, they can be just as flawed. A Canadian study, for example, as reported by USA TODAY, claims that weekend hospital admissions are more fatal. Their study was based on 3.8 million hospital admissions in Ontario from 1988 to 1997. They presumed that the greater fatality rate was due to decreased hospital staffing and inexperienced physicians. Although these appear to be logical conclusions, a closer look reveals that only 23 out of the 100 most common causes of death had higher fatality rates on weekends. Doesn"t this suggest that the other 77 most common causes of death have at least an equal chance of leading to death on weekday admissions? Furthermore, over half of the deaths were attributable to cancer, which would suggest that these patients were known to be terminally ill, and their death was unlikely due to a reduction in hospital staff.
In another study reported by USA TODAY, researchers claim that alcohol is far more of a problem than drugs for recovering substance abusers. Although it is very probable that alcohol addiction is more of a problem, the results of this study are questionable. The study consisted of a telephone survey conducted by the Alliance Project, a group consisting of substance abuse treatment centers. Not only did these researchers have a vested interest in the outcome of the study, their sample was highly biased. Results were based on 250 people who were willing to discuss their addiction. It seems logical that people would be more likely to discuss alcohol addiction since this is not an illegal activity. Substance abuse problems of the homeless are also excluded in this biased sample. .
A third study reported by USA TODAY makes the claim that pregnant women who consume eight or more cups of coffee daily, double their risk of stillbirth. This research is misleading, because not only was coffee consumption only measured once, at 16 weeks, it also did not account for several other factors.