This discussion is based on the article “Misleading Statistics & Data – News Examples For Misuse of Statistics” by Mona Lebied. Make sure to read this article before starting the discussion.

Statistics are a way of summarizing large data sets and making sense of them. Statistical results allow us to make decisions and test our preconceived opinions. While this makes statistics a powerful tool, it also means improper use can lead to misunderstanding data and making incorrect decisions. When people are trying to convince others that their arguments are the correct ones, they will use statistics to support their side. When winning is more important than the truth, they may intentionally present incorrect results and apply methodologies improperly.

The article below discusses some of the ways statistics can be used improperly to mislead other into believing one side against another. The article not only explains some common ways of doing this, it also gives some real life examples. Read the article linked above and then answer the discussion topic questions.

For this discussion, find one example where someone is misrepresenting data through improper use of statistics to support their viewpoint. Since identification of misleading statistics use can be difficult and tricky, you will use fact checking websites to find your example. Search for a case where a reliable source has identified someone misusing a statistic to mislead. Do not try to identify an example of misuse yourself.

The example you give should not simply be faulty analysis due to lack of information or skill. There should be an element of intentionally painting a misleading picture. You can find these examples in cases where there’s a conflict of viewpoints and one side is pushing their own narrative. Categorize the example you found according to the six categories described in the article. Since the hard work of identifying an attempt to mislead is already done by the fact checking source, you will be evaluated on your selection of an example and subsequent analysis of the issue.

When responding to others, please only discuss whether you think the analysis was indeed misrepresentative or not. Do not discuss the overall claim that is being made and if you disagree with it. We are only concerned about the analysis methods, not the actual assertions made. Misleading statistics are intentionally used to push narratives- we are not here to argue anything but statistics.

Please use the template below in your answers, so everyone can easily follow your answers to all the questions (using the template below is part of the requirements; you will lose points if you don’t follow the template or if you skip portions of what is being asked)

What is the title of the article? (Include a link to the article too)

What is the claim being made? Copy/paste or summarize the claim.

If it’s a visual (like a chart) then please include the chart in the post. You can usually copy paste images into the post. Include as much of the information here as possible so all readers can see what you are describing without having to visit the link and search for the problem.

What is the statistical analysis proposed to support this claim? Copy/paste their description of how they arrived at their results.

Which category does this issue fall under? Why?

Justify your selection and your identification of the issue. For example, if you say “faulty polling” you need to cite the improper wording. If you say “data fishing” you need to show they are looking for correlations without a proper hypothesis.

What would be a better analysis to evaluate the situation? Describe how to fix the faults in this analysis or suggest a different approach.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Gun-related deaths

What is the title of the article? (Include a link to the article too)

Five Thirty-Eight breaks down the more than 33,000 annual U.S. gun deaths. https://theundefeated.com/features/fivethirtyeight-breaks-down-the-more-than-33000-annual-u-s-gun-deaths/http://www.bu.edu/articles/2019/state-gun-laws-that-reduce-gun-deaths/

What is the claim being made?

“An interactive map by Five Thirty-Eight claims more than 33,000 annual gun deaths occur in the U.S.” (Five Thirty-Eight, 2017)

What is the statistical analysis proposed to support this claim? Copy/paste their description of how they arrived at their results.

“The data in this interactive graphic comes primarily from the Centers for Disease Control and Prevention’s Multiple Cause of Death database, which is derived from death certificates from all 50 states and the District of Columbia and is widely considered the most comprehensive estimate of firearm deaths. In keeping with the CDC’s practice, deaths of non-U.S. residents in the U.S. (about 50 per year) are excluded. All figures are averages from 2012 to 2014, except for police shootings of civilians, which are from 2014.” (Five Thirty-Eight, 2017).

“The “homicides” category includes deaths by both assault and legal intervention (first shootings by police officers). “Young men” are those ages 15 to 34; “women” are ages 15 and older. Because the CDC’s estimates of police shootings are unreliable, we used views from non-governmental sources. Our figure is for 2014, the first year for which such estimates are generally available. (For more on the data we used, see Carl Bialik’s story on police shootings.” (Five Thirty-Eight, 2017).

Which category does this issue fall under? Why?

On closer examination, the interactive map reveals a flawed correlation.

There is a significant bias for assuming that all deaths are as a result of gun violence. Firstly, firearm suicides make up two-thirds of the annual gun deaths in the U.S. For instance, there were 23 854 gun suicides in 2017.  But significantly, this study ignores all the gun suicides as if people aren’t as just dead as when they shoot themselves. Infact the data is presented as though all the victims were assaulted.

Why is this analysis faulty?  

There is a heated debate that gun suicide is relatively low than suicide by other means. However, there is ample evidence that people who try suicide guns rarely survive compared to other forms of suicide. Besides, there were 486 accidental deaths and 338 deaths that were unaccounted for in the study. Moreover, this statistic is false and misleading because it is assuming that the numbers are a statistic. In the last decade, there is a dramatic rise in gun violence. According to the CDC, more than 1.2 million people have faced gun-related injuries or know a victim of gun violence. For instance, the average gun deaths in 2014 averaged 12,000, but the number has risen sharply since then the latest available number is 14,542 for 2017 (Five Thirty-Eight, 2017).

What would be a better analysis to evaluate the situation? Describe how to fix the faults in this analysis or suggest a different approach.

Better analysis should include data on gun-related use such as defensive gun use, suicide deaths, and unintended deaths. Any sensible report should consider all aspects of gun use and not solely assume that all cases are gun violence-related.  There should be enough variables to show ultimate causation that limits grounds of assumption. Although the analysis is premiered on actual statistics, it prematurely cites a direct and significant correlation without breaking down the other risk factors. Therefore there is a need to analyze all the risk factors associated with firearm deaths to identify a positive and direct correlation valid enough to prove causation.

In efforts to correct the identified faultiness in the analysis, the research should include all the firearm deaths that have been purposely ignored.  All included the total number of deaths that will surpass the previously quoted figures. Overall there is a need to consider the accessibility and availability of guns as it a significant factor in gun deaths, either violence-related or other means. Furthermore, the research should examine the nature of the annual gun-related deaths.

Works Cited

Five Thirty-Eight. (2017). Five Thirty-Eight breaks down the more than 33,000 annual U.S. gun deaths. Retrieved from Five Thirty-Eight: Five Thirty-Eight breaks down the more than 33,000 annual U.S. gun deaths. https://theundefeated.com/features/fivethirtyeight-breaks-down-the-more-than-33000-annual-u-s-gun-deaths/http://www.bu.edu/articles/2019/state-gun-laws-that-reduce-gun-deaths/