Individual differences |
Methods | Statistics | Clinical | Educational | Industrial | Professional items | World psychology |
A misuse of statistics occurs when a statistical argument asserts a falsehood. In the period since statistics began to play a significant role in society, they have often been misused. In some cases, the misuse was accidental. In others, it was purposeful and for the gain of the perpetrator. When the statistical reason involved is false or misapplied, this constitutes a statistical fallacy.
The false statistics trap can be quite damaging to the quest for knowledge. For example, in medical science, correcting a falsehood may take decades and cost lives.
Misuses can be easy to fall into. Professional scientists, even mathematicians and professional statisticians, can be fooled by even some simple methods and even if they are careful to check everything. Scientists have been known to fool themselves with statistics due to lack of knowledge of probability theory and lack of standardisation of their tests.
- 1 Types of misuse
- 1.1 Discarding unfavorable data
- 1.2 Loaded questions
- 1.3 Overgeneralization
- 1.4 Biased samples
- 1.5 Misreporting or misunderstanding of estimated error
- 1.6 False causality
- 1.7 Applying group statistics to individuals
- 1.8 Statistical sleight of hand
- 1.9 Linguistically asserting unit measure when it is empirically violated
- 2 See also
- 3 References
Types of misuse
Discarding unfavorable data
In marketing terms all a company has to do to promote a neutral (useless) product is to find or conduct, for example, 20 studies with a confidence level of 95%. Even if the product is really useless, on average one of the 20 studies will show a positive effect purely by chance (this is what a 95% level of confidence means) The company will ignore the 19 inconclusive results and promote endlessly the one study that says the product/idea is good. This tactic becomes more effective the more studies there are available, because that increases the likelihood that there will be a large number of studies that support the product's effectiveness. For example, at a 95% confidence level, if there are 200 studies then on average 10 of them will show an effect. Organisations that do not publish every study they carry out, such as tobacco companies denying a link between smoking and cancer, or miracle pill vendors, are likely to use this tactic.
Another common technique is to perform a study that tests a large number of dependent (response) variables at the same time. For example, a study testing the effect of a medical treatment might use as dependent variables the probability of survival, the average number of days spent in the hospital, the patient's self-reported level of pain, etc. This also increases the likelihood that at least one of the variables will by chance show a correlation with the independent (explanatory) variable.
Often, people or organizations that perform multiple studies to attempt to get a certain result will either classify the data from the studies that didn't produce the desired outcome as "secret" or claim that the data was lost due to computer failure in order to prevent their fraud from being discovered.
The answers to surveys can often be manipulated by wording the question in such a way as to induce a prevalence towards a certain answer from the respondent. For example, in polling support for a war, the questions:
- Do you support the attempt by (the war-making country) to bring freedom and democracy to other places in the world?
- Do you support the unprovoked military action by (the war-making country)?
will likely result in data skewed in different directions, although they are both polling about the support for the war.
Another way to do this is to precede the question by information that supports the "desired" answer. For example, more people will likely answer "yes" to the question "Given the increasing burden of taxes on middle-class families, do you support cuts in income tax?" than to the question "Considering the rising federal budget deficit and the desperate need for more revenue, do you support cuts in income tax?"
If you have a statistic saying 100% of apples are red in summer, and then publish "All apples are red", you will be overgeneralizing because you only looked at apples in summertime and are using that data to make inferences about apples in all seasons. Oftentimes, the overgeneralization is made not by the original researcher, but by others interpreting the data. Continuing with the previous example, if on a TV show you say "All apples are red in summer" many people will not remember you said "in summer" if asked weeks later.
With a subject on which the general public has no personal knowledge of, you can fool a lot of people. For example you can say on TV "Most autistics are hopelessly incurable if raised without parents or normal education" and many people will only remember the first part of the claim, "Most autistics are hopelessly incurable". This problem is especially prevalent on TV, where talk show hosts interview one individual as representative of a whole class of people.
- Main article: bias (statistics)
Misreporting or misunderstanding of estimated error
If a research team wants to know how 300 million people feel about a certain topic, it would be impractical to ask all of them. However, if the team picks a random sample of about 1000 people, they can be fairly certain that the results given by this group are representative of what the larger group would have said if they had all been asked. This is called the law of large numbers.
This confidence can actually be quantified by the central limit theorem and other mathematical results. Confidence is expressed as a probability of the true result (for the larger group) being within a certain range of the estimate (the figure for the smaller group). This is the "plus or minus" figure often quoted for statistical surveys. The probability part of the confidence level is usually not mentioned; if so, it is assumed to be a standard number like 95%.
The two numbers are related. If a survey has an estimated error of ±5% at 95% confidence, it might have an estimated error of ±6.6% at 99% confidence. The larger the sample, the smaller the estimated error at a given confidence level.
Most people assume, because the confidence figure is omitted, that there is a 100% certainty that the true result is within the estimated error. This is not mathematically correct.
Many people may not realize that the randomness of the sample is very important. In practice, many opinion polls are conducted by phone, which distorts the sample due in several ways, including exclusion of people who do not have phones, favoring people who have more than one phone, favoring people who are willing to participate in a phone survey over those who refuse, etc. Non-random sampling makes the estimated error unreliable.
On the other hand, many people consider that statistics are inherently unreliable because not everybody is called, or because they themselves are never polled. Many people think that it is impossible to get data on the opinion of dozens of millions of people by just polling a few thousands. This is also inaccurate. A poll with perfect unbiased sampling and truthful answers has a mathematically determined margin of error, which only depends on the number of people polled.
The problems mentioned above apply to all statistical experiments, not just population surveys.
There are also many other measurement problems in population surveys.
- Main article: Correlation implies causation (logical fallacy)
When a statistical test shows a correlation between A and B, there are usually four possibilities:
- A causes B.
- B causes A.
- A and B are both caused by a third factor, C.
- The observed correlation was due purely to chance.
The fourth possibility can be eliminated by statistical tests that can calculate the probability that the correlation observed would be as large as it is just by chance. However, even if that possibility is eliminated, there are still the three others.
If the number of people buying ice cream at the beach is statistically related to the number of people who drown at the beach, then nobody would claim ice cream causes drowning because it's obvious that it isn't so. (In this case, both drowning and ice cream buying are clearly related by a third factor: the number of people at the beach).
This fallacy can be used, for example, to prove that exposure to a chemical causes cancer. Replace "number of people buying ice cream" to "number of people exposed to chemical X", and "number of people who drown" with "number of people who get cancer", and many people will believe you. In such a situation, there may be a statistical correlation even if there is no real effect. For example, if there is a perception that the chemical is "dangerous" (even if it really isn't) property values in the area will decrease, which will entice more low-income families to move to that area. If low-income families are more likely to get cancer than high-income families (this can happen for many reasons, such as a poorer diet or less access to medical care) then rates of cancer will go up, even though the chemical itself is not dangerous. It is believed that this is exactly what happened with some of the early studies showing a link between EMF fields from power lines and cancer.
In well-designed studies, the effect of false causality can be eliminated by assigning some people into a "treatment group" and some people into a "control group" at random, and giving the treatment group the treatment and not giving the control group the treatment. In the above example, a researcher might expose one group of people to chemical X and leave a second group unexposed. If the first group had higher cancer rates, the researcher knows that there is no third factor that affected whether a person was exposed because he controlled who was exposed or not, and he assigned people to the exposed and non-exposed groups at random. However, in many applications, actually doing an "experiment" in this way is either prohibitively expensive, infeasible, unethical, illegal, or downright impossible. (For example, it is highly unlikely that an IRB would accept an experiment that involved intentionally exposing people to a dangerous substance in order to test its toxicity.)
Applying group statistics to individuals
- Main article: Ecological fallacy
Statistical sleight of hand
- Main article: Statistical special pleading
Missing important caveats.
Linguistically asserting unit measure when it is empirically violated
Unit measure is an axiom of probability theory which states that, when an event is certain to occur, its probability is 1. This axiom is consistent with the empirical world, if the relation from a set of events that are certain to occur to a set of physical objects is one-to-one, but not otherwise. In the latter case, unit measure is scientifically invalidated.
Christensen and Reichert (1976), Oldberg and Christensen (1995) and Oldberg (2005) report observations of systems in which the relation is not one-to-one. A result of the lack of one-to-one-ness is that the following elements of statistical terminology are not defined for the associated systems: a) "population", b) "sampling unit", c) "sample", d)"probability", e) any term that assumes probability theory. A misuse of statistics arises when any of these terms are used in reference to a system that lacks one-to-one-ness, for unit measure is linguistically asserted and empirically violated. Oldberg and Christensen (1995) and Oldberg (2005) report observations of this type of misuse.
- Christensen, R. and T. Reichert, 1976 "Unit Measure Violations in Pattern Recognition, Ambiguity and Irrelevancy," Pattern Recognition, vol. 4, pp. 239-245. Pergamon Press.
- Hooke, R., 1983, How to tell the liars from the statisticians; Marcel Dekker, Inc., New York, NY.
- Jaffe, A.J. and H.F. Spirer, 1987, Misused Statistics; Marcel Dekker, Inc., New York, NY.
- Campbell, S.K., 1974, Flaws and Fallacies in Statistical Thinking; Prentice Hall, Inc., Englewood Cliffs, NJ.
- Oldberg, T., "An Ethical Problem in the Statistics of Defect Detection Test Reliability," 2005, Speech to the Golden Gate Chapter of the American Society for Nondestructive Testing. Published on the Web by ndt.net at http://www.ndt.net/article/v10n05/oldberg/oldberg.htm.
- Oldberg, T. and R. Christensen, 1995, "Erratic Measure" in NDE for the Energy Industry 1995; The American Society of Mechanical Engineers, New York, NY. Republished on the Web by ndt.net at http://www.ndt.net/article/v04n05/oldberg/oldberg.htm.