How Would Jesus Calculate Standard Deviation?
“ChristiaNet Poll Finds That Evangelicals Are Addicted to Porn,” screamed a Monday headline on Market Wire.There’s just no way I can simply sail past a headline like that, even on a busy Monday morning.
The news bulletin (which appears to be little more than a press release issued by ChristiaNet.com, a destination proclaiming itself “the world’s most visited Christian website”) continued beneath the headline to set out on a not-so-objective footing, ultimately reporting that a recent opinion poll reveals an epidemic of porn addiction among America’s “evangelical” Christian community.
Citing 1000 responses to an 11-question online opinion poll, the results of which ChristiaNet.com reportedly partnered with Second Glance Ministries in evaluating, ChristiaNet.com makes some astounding claims.
“The poll results indicate that 50-percent of all Christian men and 20-percent of all Christian women are addicted to pornography,” said Clay Jones, founder and president of Second Glance Ministries.
Um… yeah, that just sounds right, intuitively, doesn’t it?
Jones, who told Market Wire that his ministries’ objectives include “providing people with information which will enable them to fully understand the impact of today’s societal issues,” went on to claim that among respondents:
• 60-percent of women “admitted to having significant struggles with lust”
• 40-percent of women “admitted to being involved in sexual sin in the past year”
• 20-percent of “church-going female participants struggle with looking at pornography on an ongoing basis”
“There have been dynamic paradigm shifts in the behavior of Christians over the last four years,” said Jones. “Technology has allowed pornography to flood the market place beyond a controllable level.”
Bill Cooper, president of ChristiaNet.com, echoed Jones’ claims, saying that his group has observed “escalation to the [porn addiction] problem in both men and women who regularly attend church.”
After reading the article, I wondered how ChristiaNet.com’s survey might be received by their peers at Morality In Media (MIM), who reported last month that a poll commissioned by their group indicated that nearly three-quarters of all Americans consider viewing pornographic videos or websites to be “morally unacceptable.”
The results of the MIM poll, a telephone survey of 997 adults (18 and over) within the United States last month, were heralded by MIM President Robert W. Peters as evidence that Americans don’t approve of porn, despite growing revenues in the adult entertainment sector and the “mainstreaming” of pornography in recent decades.
“Those who defend pornography, whether in court or in the court of public opinion, point to the proliferation of this sordid material as ‘proof’ either that everyone is viewing it or that people no longer deem pornography unacceptable,” said Peters.
On the contrary, said Peters, “much if not most pornography is consumed by a relatively small percentage of males who are hooked on it.”
Plus, Roberts says; “just because a person, whether out of curiosity or at a weak moment or for a period of time, views pornography does not mean he has become a devotee of it.”
Perhaps Roberts wouldn’t take issue with the ChristiaNet.com survey, after all, despite the fact that the results ChristiaNet.com claims seem to contradict the MIM conclusion that porn consumption is largely limited to a “small percentage of males who are hooked on it” – after all, Roberts himself says that even a porn addict can “disapprove” of porn.
“(E)ven among long-term users,” Roberts said, “not all approve of their own behavior. Many addicts hate what they do.”
So, American adults disapprove of porn, but they are also addicted to it? These are both ostensibly good Christian men and we certainly don’t want to suggest they are intentionally promoting statistics that they know to be false.
Perhaps neither group is intentionally misleading anyone – perhaps they’re just lousy researchers and/or sub-par data analysts?
Statisticians (and one would hope, just about any elementary school graduate) will tell you that any manner of “opinion poll” survey is highly limited when it comes to producing reliable statistical data of any kind.
The wording of questions in opinion polls can be, and frequently is, structured by agenda-driven pollsters to elicit the responses that best match their own arguments, conclusions, and positions.
Beyond the phraseology used by pollsters, in order to arrive at anything like a “representative sample,” something potentially useful in projecting results beyond the surveyed population, pollsters must at least ensure that the sample itself is derived randomly.
A survey conducted by a Christian website is very unlikely to produce a random responding population. There’s also a noticeable lack of means by which one could ascertain whether an apparently random sample obtained through an online survey is, in fact, random.
Thus, the methodology of the ChristiaNet.com survey was flawed, fatally so, from the word go.
The MIM poll, while superior in methodology to the ChristiaNet.com survey, should still be regarded with substantial skepticism, especially given that it was a vocally anti-porn activist group that commissioned the survey and structured the questions therein. Beyond natural suspicion of the surveyor, though, there are more fundamental weaknesses in the MIM survey.
In its press releases and interviews concerning the survey, the MIM does not specify the geographic distribution of its respondents (saying only that they are “within the United States”). The MIM also doesn’t indicate how (or if) it defined “morally acceptable” within the survey questions, or how/if “pornography” itself was defined, for that matter.
It does add somewhat to the credibility of the MIM survey that MIM employed Harris Interactive to conduct the polling, inasmuch as Harris presumably followed its usual protocols for reducing the survey’s “margin of error.”
It is important to understand, however, what “margin of error” means – and what it doesn’t mean – in the context of survey results. This is a crucial point generally avoided in media reports of statistical data, for the simple reason that most journalists have no clue what it means, but happily throw the phrase “margin of error” around, nonetheless.
Assuming a truly random population of Americans is selected, a survey of 997 Americans does indeed provide a “95-percent probability that the overall results have a sampling error of +/- 3 percentage points,” as MIM claims in its press release.
All the above means, however, is that there’s a 95-percent likelihood another random survey would yield a variation in responses to the same questions of greater than 3-percent from the results of the MIM survey.
When it comes to the issue of what survey responses actually indicate about public opinion, the operative question in my mind isn’t “What does the survey indicate about how people respond to these questions?”
Some questions are simple and a very small number of possible responses can be presented, accordingly. The public’s preference between candidates in a race for political office, for example, can be gauged fairly accurately with a random sample of one or two thousand people.
So, which porn survey would Jesus endorse?
Well, unless Jesus was pretty damn gullible – and really bad at data analysis – he might well have judged both to be Statistical Apostasy of the worst kind.