The Misuse of Statistics:

Common ways of concealing the truth

Statistics, if correctly applied, form the basis of rational decision making across all fields where data is collected and quantified, If incorrectly applied, statistics lead to errors in decision making and downright deception ((3) (PDF) The Misuse of Statistics: Concepts, Tools, and a Research Agenda (researchgate.net)This note  explores  the various ways in which statistics can be misused and the consequences if they are.

Selective Presentation of Data

One common misuse of statistics is the selective presentation of data to argue in favour of a particular conclusion. By cherry-picking data points that support a particular argument while ignoring those that don’t, individuals can create a misleading narrative  or (through context omission) support an incorrect hypothesis. A newspaper headline once declared “Men go to supporting centres to meet women” based on a survey of men at a shopping centre. The data did show this but, but instead of highlighting some salacious activity by men at shopping centres it simply confirmed that the men surveyed were going to meet their wives as part of a weekly shopping activity.

Misleading Graphs and Charts

Graphs and charts are visual representations of data that can be easily manipulated to deceive the viewer. For example, altering the scale of a graph can exaggerate trends, and omitting baseline values can make differences appear more significant than they are. Such distortions can lead to false conclusions and misinformed decisions.

Source bad visualisations (https://www.tumblr.com/badvisualisations/184828044636/1-as-a-rule-the-y-axis-of-a-bar-graph-needs-to?

As a rule, the Y axis of a bar graph needs to start at zero. This violates that. Makes it seem like Indian women are one-fifth the size of Latvian women.  As well why are the figures scaled on the X-axis as well? Adds no information

Improper Use of Correlation and Causation

A common statistical error is confusing correlation with causation. Just because two variables move together does not mean one causes the other. Misinterpreting these relationships can lead to erroneous beliefs, such as assuming a superstition is true because it coincides with an outcome.

This graph shows close graphical correlation between UFO sightings in the US state of South Carolina and the total number of UFO sightings and was used in the satirical journal . The data on these things had strong diagnostic support with the data over that period showing

• Correlation r = 0.9171843(Pearson correlation coefficient) The highest value possible is 1 meaning the data moved in the same direction 92 % (approximate) of the time.
  Regression  Correlation  of and an  r² = 0.8412271 (Coefficient of determination)

This means 84.1% of the change in the one variable (i.e., Total Number of Successful Mount Everest Climbs) is predictable based on the change in the other (i.e., UFO sightings in South Carolina) over the 37 years from 1975 through 2011. The problem with this analysis was that there was no underlying or logical reason for these two variables to be related. An important rule of thumb when using statistics to show correlation or causation is reasonableness and looking for some other variable(s) that cause both and determine the apparent association.  If you can’t think of a logical explanation , there probably isn’t one

Overgeneralization from Small Samples

Drawing broad conclusions from small or non-representative samples is another misuse of statistics. This can occur in scientific studies or surveys where the sample size is too small to be indicative of the larger population. Overgeneralization can lead to stereotypes and policies that do not reflect the true nature of the issue at hand.

Ignoring Margin of Error

The margin of error is a crucial concept in statistics that reflects the uncertainty in estimates. Ignoring this can give a false sense of precision. For instance, in election polling, failing to account for the margin of error can lead to incorrect predictions and public mistrust in the polling process.

Consequences of Misuse

The misuse of statistics can have far-reaching consequences. It can undermine public trust in institutions, lead to poor policy decisions, and even cause financial or health-related harm. In a world increasingly driven by data, it is vital to maintain ethical standards in the presentation and interpretation of statistical information.

Conclusion

Statistics are invaluable in making informed decisions where sufficient data allows meaningful to be drawn. Where this isn’t the case,  statistics can be used to confuse , distort sand deceive. In interpreting statistics be ware of sample size, tests of significance of variables and the presence of omitted variables that in reality are the prime cause of the association.