Statistical literacy

From Wikipedia, the free encyclopedia

Statistical literacy is a term used to describe an individual's or group's ability to understand statistics. Statistical literacy is considered by many to be necessary for citizens to understand material presented in publications such as newspapers, television and the internet and is an area of growing importance in statistics education. Numeracy is a prerequisite to being statistically literate. Many official statistical agencies such as Statistics Canada have programs to educate students in schools about the nature of statistics.

One could consider thinking critically and thus critical pedagogy to be broader domains that should contain statistical literacy, as thinking statistically is one key skill of a statistical education. Thinking statistically would be similar to other professional skills such as thinking like an engineer or doctor. But statistical literacy generally is known to many professions not just the professional statistician.

Experiments in the sciences and business models and reports use statistics and people involved in these fields generally have studied the meaning of statistical quantities, such as averages and standard deviation. In fact, many colleges and universities require an introductory course in statistics as part of a professional program. This can be a barrier for some to graduation. Statistical anxiety is almost the same as math anxiety in this context. Poorly constructed statistics in the media do not help this situation in the schools although the example of a bad statistic is a good motivator for learning the right way of doing statistics. The books referenced below have such examples.

Each day people are inundated with statistical information from advertisements ("4 out of 5 dentists recommend"), news reports ("opinion poll show the incumbent leading by four points"), and even general conversation ("half the time I don't know what you're talking about"). Experts and advocates often use numerical claims to bolster their arguments, and statistical literacy is a necessary skill to help one decide what experts mean and which advocates to believe. This is important because statistics can be made to produce lies and misrepresentations of data that may seem valid. The aim of statistical literacy proponents is to improve the public understanding of numbers and figures

Numbers related to a countries people, such as labour force statistics, gross national product, population size, literacy levels, age of workers, gender of workers, and other detailed statistics help define both a country and the original meaning of the word statistics... state metrics. These values while "just numbers" and in some sense considered objective neutral numbers will cause opinions to form. These opinions about a country or an individuals position within that country, not to mention the opinions we all hold of others in our own country and others in yet other countries can not be considered to value free or "just numbers" once they have context.

Further, because statistics are used to adjust social programs, investments, and rather complex emotions and social interactions that result from expressing or creating opinions, statistics need to be understood by many.

Also many statistics are calculations or estimates of real numbers. As hinted to above, real numbers may be real numbers for one advocate but for another advocate the wrong real numbers to be estimating. Understanding that these estimates are predictions requires some understandings of probabilities. So opinions are not just measured as in polls but also created by reading and interpreting statistics and probabilities are also opinions in numerical form of the likelihood of finding a good prediction.

Results of opinion polling are often cited by news organizations, but the quality of such polls varies considerably. Some understanding of the statistical technique of sampling is necessary in order to be able to correctly interpret polling results. Sample sizes may be too small to draw meaningful conclusions, and samples may be biased. The Alexa Internet web traffic reports, for example, are known to be biased for several reasons, one of which is that their toolbar only works with the Internet Explorer browser [1]. The wording of a poll question may introduce a bias, and thus can even be used intentionally to produce a biased result. Good polls use large samples and unbiased techniques, with much time and effort being spent in the design of the questions and polling strategy. Statistical literacy is necessary to understand what makes a poll trustworthy and to properly weigh the value of poll results and conclusions.

A problem also occurs with literacy because of the work of statisticians. The legibility of numerical tables is an example given early in the recent book by Richard M. Heiberger & Burt Holland, Statistical Analysis and Data Display ( New York, N.Y.: Springer, 2004). In their example, rather than an incorrect, confusing, collection of numbers in misaligned columns, the statistician must present results legibly by lining up the decimal points so the visual presentation is more organised [Heiberger]. This would be an example of the statistician increasing statistical literacy rather than making matters worse.