April 30, 2007
The use (or misuse) of statistic
In response to Virginia Tech’s shooting event, I heard the following statistic during a sunday sermon. The statistic is based on a survey conducted on currently or previously imprisoned juveniles. I don’t remember the exact details of the statistic, but it went something like this:
- Out of approximately 2000 respondents, only 5% believed in the existence of God,
- only 3% have ever stepped in a church, and
- only 1% currently attend church
The pastor used the above numbers to imply that, had the church done a better job in attracting young people to God, fewer youngster would end up in prison.
Although I have no doubt in the conclusion that God can change young deliquent’s life, I do have my doubts on the supporting statistic. Somehow, I find the misuse of statistic quite widespread in radio shows and some popular magazines.
The first problem with the given statistic is that we are not told what is the average number of young people who believe in God but have NEVER been imprisoned. For the conclusion to be true, the number of young people who believe in God outside of prison should be far greater than 5%. But is this really true? I personally wouldn’t be surprised if only 1 out of 20 average young people believe in God nowadays.
The second problem that I see with this example is that it assumes the interviewed people kept their belief (or lack of it) during they stay in prison. But what if certain people believed in God while in prison? Or, conversely, what if certain people lost their faith in prison? Either situation would directly skew the results in one way or another, effectively weaking the conclusion.
And, finally, even if the previous two assumptions were true, it still doesn’t prove (statistically) that attracting young people to church will keep them out of prison. The numbers merely suggest that there is correlation between people who don’t believe in God and people who end up in prison. We can’t establish a cause-and-effect situation with only the numbers that were given.


Comments(2)