Misuse of statistics

This article may require cleanup to meet Wikipedia's quality standards. The specific problem is: this article is badly written, often unclear. Please help improve this article if you can. (November 2014) (Learn how and when to remove this message)

The misuse of Statistics can trick the observer who does not understand them into believing something other than what the data shows or what is really 'true'. That is, a misuse of statistics occurs when an argument uses statistics to assert a falsehood. In some cases, the misuse may be accidental. In others, it is purposeful and for the gain of the perpetrator. Often, when the statistics are false or misapplied, this constitutes a statistical fallacy.

The consequences of such misinterpretations can be quite severe. For example, in medical science, correcting a falsehood may take decades and cost lives.

Misuses can be easy to fall into, and often statistical tests have assumptions which cannot be met in reality, such as the independence of observations. For instance, if we are studying kids in a classroom, they are not independent because they have the same teacher, go to the same school, probably live in the same area of the country, etc. In such cases, these assumptions can often be ignored, but often they should not be. Professional scientists, mathematicians and even professional statisticians, can be fooled by even some simple methods, even if they are careful to check everything.

A recent example of this is work on scientific fraud that looks at image duplication, but then moves to the paper or journal level to make their claim. For instance, if a team finds that .1% of images are duplicated (1 of 1000), they can report this, which is not that exciting. Conversely, they can summarize their data to the paper or even journal level, thus inflating their statistics, because each paper and journal contains many images. For instance, they can report e.g., that 5% of papers have such an image in them, if each paper contains e.g., 50 images, or even to the journal level for instance saying that 90% of journals had at least 1 duplicated image in them (if each journal contains 10,000 images). This practice is similar to trying to understand for instance how many highschoolers cannot read in e.g., the United States, but then summarizing the data to the school or state level if you find one student that cannot read in those units. Such manipulations occur because scientist's jobs and funding often depend on the problem they are working on being important, thus in our case the researcher can suggest that image manipulation or inability to read is a far larger problem than it probably is.

Scientists have been known to fool themselves with statistics due to lack of knowledge of probability theory and lack of standardization of their tests.