EDITORIAL – 95 percent of all statistics are skewed
October 26, 2005
Sometimes, statistics can be misleading. A recent study that links religion to wealth is a… Sometimes, statistics can be misleading. A recent study that links religion to wealth is a prime example of how we must carefully scrutinize all the numbers used to affirm alleged causation or correlation.
Apparently, Jonathan Gruber, of the economics department of Massachusetts Institute of Technology, asserts that going to church on a regular basis can lead to an increase in household income by 9.1 percent.
Gruber also notes increases in the level of education and marriage, and notes a decrease in having a handicap, being on welfare and getting a divorce. His report goes on to explain this by stating that those who attend church regularly are more successful because they are less stressed about the routine inhibitors of success.
Those who read the headlines that state “Religion linked to wealth” can be easily convinced with the empirical nature and persuasiveness of Gruber’s report, “Religious Market Structure, Religious Participation and Outcomes: Is Religion Good for You?”
This is not to say that his suspicions and his data are completely unfounded. Other findings that precede his strongly suggest those who are terminally ill fare better when they have a support system or faith – thus affirming the power of religion.
But the scrutiny in his findings dwells in his methods of research and our own naivete. Beginning with the latter, we must first understand that just because there is a correlation doesn’t mean there is causation. In other words, the fact that religion is a factor in acquiring success doesn’t mean that going to church will increase your cash flow.
The world is not a vacuum and we can’t trivialize other factors that come into play when we make conclusions or sweeping generalizations. For example, those who attend sizable churches with sizable congregations are afforded networking opportunities that leave those sleeping in on a Sunday morning at a disadvantage.
Let’s observe the fellow churchgoer who just got laid off. In the middle of the service brother Matthew walks his distraught, downtrodden self to the pulpit to give his “testimony” to a congregation of hundreds who also happen to be spread out among myriad occupations and careers. While prayer can definitely be a factor in the job Matthew is likely to be offered in a matter of minutes, the makeup of his church is definitely a considerable factor.
Taking a closer look at the study at hand, it seems this gentleman from MIT missed the mark. Gruber’s study focuses significantly on non-Hispanic whites who were in their mid 20s or older.
In doing this he completely eliminated Hispanics and blacks, who not only happen to have significant numbers in the religious community but also densely populate the middle- to lower-class sections of the economic pyramid significantly more than their white counterparts. To boot, in Gruber’s categories of religion he considered Catholics, Jews and liberal, moderate and conservative Protestants but excluded those of the Muslim faith.
So the next time you read a headline or a lead in a story with an impressive statistic, do your homework. More often than not there are flaws that can shed light on possible factors capable of swaying your opinion.