3 Ways the Media Uses Statistics to Manipulate People Video and Transcript
3 Ways the Media Uses Statistics to Manipulate
Number 3: Percent Increase without Relative Context
When those in the media want to scare people into thinking some issue is drastically increasing, they will often use a percent increase statistic without context.
To understand what a percent increase is, let’s go over the basics. A percent increase is the comparison of how much something has gone up from an original state.
For example, if there was one murder that took place in a certain county last year, and this year, 2 murders took place, the percent increase would be 100%, calculated by subtracting this year’s murder count from last year’s murder count, dividing that result by last year’s murder count, multiplying the result, in this case, a 1, by 100, and adding a percent sign.
In this way, the media can make it sound like there is a scary increase of murders by saying there was a 100% increase in murder, while the actual nominal count of going from 1, to 2, while tragic, is not that drastic of a number, especially if the population in the county is over 100,000 people.
So, whenever you see reporters talking about a percent increase to make a situation seem scary, be sure to follow up on exactly how much of a number increase that is.
Often, you’ll find that the actual increase is not as scary as they are making it out to be.
Number 2: Measure of Central Tendency - Median versus Mean
Measures of central tendency are figures used to state or estimate what a seemingly typical experience is. Whether median or mean is used can drastically change how a situation is interpreted.
For a simple example, let’s look at 5 people and their incomes.
Alfred made $30,000 last year.
Betty made $35,000 last year.
Charles made $40,000 last year.
Diana made $45,000 last year.
And Elaine made $1,000,000 last year.
To get the mean income, we would add up all their incomes and divide them by the number of people in this group. This amounts to having one million, one hundred fifty thousand dollars divided across 5 people, for a result of $230,000 as the mean.
Now, does $230,000 represent the typical income of the 5 people? Clearly, not, as 4 out of the 5 people make nowhere close to $230,000. Elaine’s $1,000,000 income skews the data so much that the average does not really represent the typical earner’s money.
In this case, taking the median, or the middle value from lowest to highest earnings, produces a much clearer picture of the typical income.
As Charles’ $40,000 income is the 3rd one in with two others to the left and right of him, his income is the median and is much closer to most of the others’ income in this group.
As you can see, it is important to note which measure of central tendency is being used in the media, as the media may use mean or median to hide key facts about the typical experience of a certain group.
Number 1: Using Correlation to Imply Causation
Corporate media heads often like to pair two events together to make it appear as if one thing caused another, whether that’s gun ownership and homicides, or living longer by moving to a big city in California.
The key takeaway is that just because two events happened at the same time and seem to follow a pattern, that does not, in and of itself, present causation without observational study of what specifically lead from one thing to another to rule out other possible factors.
For a good example of how the media uses correlation wrongly, especially considering other factors, let’s look at BJ Campbell’s article: Everybody’s Lying About the Link Between Gun Ownership and Homicide.
Campbell notes in his article that many gun-grabbing news outlets like to make headlines that America has a “gun problem,” while citing studies on correlations about homicide, guns, and other factors.
Campbell points to one often-cited study published at the American Journal for Public Health titled, “The Relationship Between Gun Ownership and Firearm Homicide Rates in the United States, 1981–2010.”
He notes that, in this study, the authors found that for each 1 percentage point increase in proportion of household gun ownership [via gun suicide proxy], firearm homicide rate increased by 0.9%.
In addition, for each 1 percentage point increase in proportion of Black population, firearm homicide rate increased by 5.2%.
For each 0.01 increase in Gini coefficient, that is, income inequality, firearm homicide rate increased by 4.6%.
Thinking about these comparisons, being black and poor had about a 5 times higher factor for homicide increase over household gun ownership.
So do outlets like Mother Jones, Vox, or CNN suggest that there is a “poor black problem” as those categories hold a much higher correlation than just household gun ownership?
Of course not.
And they shouldn’t.
Because in all cases of gun ownership, poverty, and ethnicity, correlation is not causation.
As Campbell notes, urbanized areas where there are other environmental and social factors can be the reason for a gun purchase in the first place.
And the number of minorities who peacefully own guns, even those who are poor, far outnumbers those who use guns for violence.
So, if Everytown for Gun Safety or Moms Demand Action state that “We have gun problem,” they would be making the exact same correlation error as if they said, “We have a poor minority problem.”
As you can see, it is crucial to carefully consider how the media frames two events coming together.
If they cannot show, observationally, that one thing causes another by eliminating other factors, be careful to check the underlying data.
Often, there is a spurious correlation being wrongly promoted as if it were causation.
<bEverybody’s Lying About the Link Between Gun Ownership and Homicide