Lies and Statistics

• Trends
• August 20, 2014
• by Merrie Spaeth

Lies, damned lies and you know what? This week I came across an interesting example of how statistics can be fundamentally misleading but can shape a debate when they are repeated over and over.

Perhaps you’re following the national discussion and debate about sexual assaults on college campuses. Legislation has been proposed in Congress. The Department of Education has proposed extensive regulations on how complaints should be investigated, adjudicated and reported. The statistic driving the urgency is that “one in five students will be a victim of a sexual assault in college.” That figure shocks the listener and creates a sense of urgency.

The only problem is that the statistic is misleading to the point of a lie.

The “study” that generated the number looked at only two schools and there were a significant percentage of participants who didn’t respond. Additionally, the survey language was so vague as to be useless. While the word “assault” creates the image of violence and rape, the survey actually allowed for everything on the spectrum, including unwanted touching. Nor did it deal with the major dynamic driving the discussion: alcohol. Almost no male students realized that the use of alcohol invalidates the concept of mutual consent. Few female students recognized the extent to which alcohol loosens inhibitions among both males and females.

Thoughtful people are rightfully skeptical about regulations written by and administered by Washington bureaucrats. Aside from the merits of the debate, the example illustrates how the media and regulators latch onto a statistic that appears real and escalate it into policy with the potential to impact thousands of institutions and millions of lives.