The use of statistics to sway opinion has become an integral part of the national conversation. It is an effective method to give an authoritative air to one’s views. The incentives are significant for those compiling or interpreting data to skew conclusions to bolster their own prejudices. While this type of manipulation can range from the simplistic to the sophisticated, these pseudo-scientific efforts can at times have the desired effect. Since accessing and investigating source material is often time-consuming and unpleasant, many assertions hang in the air unchallenged, left unexamined by opposing viewpoints.
One of the most famous examples of how high-profile statistical measurements can be woefully inaccurate was the widespread prediction of a Hillary Clinton win. Poll after poll pointed favorably to a new Democratic White House, resulting in deep psychological shock when these predictions did not pan out. In order to preserve their integrity, some polling organizations cast the inaccuracies in the most unflattering light against Donald Trump. Arguing that less educated voters were ‘a key demographic for Trump’ and ‘are consistently hard for pollsters to reach’, the prestigious Pew Research center had great difficulty identifying reasons for the widespread errors. Investigative reporter Sharyl Attkisson posited a more cynical analysis. In her excellent book ‘The Smear’ Miss Attkisson shares her experiences poring through methodology disclosures and finding widespread dishonest practices throughout the mainstream media. These include oversampling of Democrats and ignoring favorable trends toward Trump.
In her book ‘Adios America’, Ann Coulter recounts the difficultly of even finding statistics about illegal immigrant crime. According to her experiences “every year except 1925, Hispanics are counted as ‘White’.” Another example of measurements downplaying the numbers include comparing immigrant crime to a sampling of American criminals, rather than to the general American population. Oversampling of native-born American Blacks and Latinos we also used for comparisons – groups that themselves are (a) not representative of the overall American demographic and (b) already have higher crime rates than the general society. Of course, if a case could be made that immigration – both legal and illegal – comes with increased rates of crime, that conclusion could affect national elections. Otherwise, famous Mexican-American journalist Jorge Ramos (who the New York Times called ‘the Walter Cronkite of Latino America’) can confidently claim that “Immigrants are less likely to be criminals than US Citizens” and that “Immigrant crime rates are lower than born Americans.”
Another example of the manipulation of statistics includes the campus rape culture ‘epidemic’. The liberal Huffington Post website has informed its readers of the staggering number – “about 1 in 5 women may experience sexual assault at college” it announced. At some schools the rate of sexual assault was as high as 1 in 2 the article claimed. The interesting twist regarding these statistics is the change of the word ‘rape’ to a new term ‘sexual assault’. A change of definition makes interpretation of numbers more subjective. More importantly, other more comprehensive statistics from the Justice Department suggest a much lower number of 1 in 52 college women who have been victims. Despite conflicting claims, the 1 in 5 number has become the basis for a slew of government and campus policies and is difficult to address because of the serious nature of the crime and an unwillingness to appear insensitive to the plight of women.
There are very strong incentives for both governmental and private agencies to have alarming numbers for identifiable crises as this can be used as justification to increase government funding, staff and other resources to the areas affected. Any government administration may appoint department heads that can mandate or at least influence the fact-collecting process. Not only does the government decide what data to keep, but it also defines and collates that data. The state may be perfectly comfortable ensuring that the numbers are difficult to understand and that classification of data points shroud underlying truths. Besides protecting their own budgets, other problems include inertia or simple incompetence in the system of compiling information. This indirect institutional bias works against any outside parties who are not sold on the government’s claims. Economist Thomas Sowell has written several poignant columns regarding the misleading use of statistics in general because of the institutional imperatives involved. For decades, Journalist Michael Fumento has written commentary on public-issue areas where manipulation of the general public was in the interest of various pressure groups. While this blog does not take a stance on the validity of these writings, it can certainly be argued that wherever statistics reaffirm conclusions that the institutions who generate those numbers are concerned, there should be outside parties verifying the research involved.
As a general rule, it seems that having an accurate view of reality as possible, given the resources available, is in our interest – at least where developing political strategy is concerned. It is not inconceivable that there would be instances where right-wing groups would paint a picture most in accordance with their objectives, but these incidents would have to be assessed on a case by case basis as to their significance in the broader society. Employing a group of dedicated statisticians who would scrutinize at least the most famous yet suspicious claims may be a useful addition to a more robust ideological arsenal.