Introducing Some Ideas in Statistics

Introduction to some ideas of statistics

SampleTesting

Destructive testing is sometimes used to test the ultimate life of a product or design.  In the process of testing, the item is destroyed, so valuable information is gained, but obviously not every unit can be tested, so we rely on sample testing.  How many units need to be tested, and how frequently during the production cycle, are questions statistics addresses.

Sampling can also be used when it is not feasible, or cost effective, to measure every member of a set.  Sampling to figure out the size of fish in a lake is an example.  How much sleep does the average U.S. high school student get per night?

Surveys can reveal a lot of useful information, but care has to be taken to make sure that the data is not tainted by any biases. Biases can affect what questions are asked and the order of the questions which can often affect the results.  A big question in statistical studies is, who gets surveyed?  How the participants are selected can affect the accuracy and outcome of the survey.

Examples of biases are:

Age bias, religious bias, political bias, racial bias, and the list goes on.  Biases can affect the outcome of studies and often show up in news reporting of statistical studies, which might be highlighted differently by different parties with different outlooks. Bias can be conscious or unconscious.

Data presentation

Data is often presented in graphical form so that it is easier to see the big picture at a glance. Which graph is best?  the scale and which part of the data is to be highlighted are all part of data presentation.

Two variables can be shown to correlate, but that is not the same as showing a causal relationship.  Are the two variables changing in sync because one is causing the other or because of some other factor?  As we say, “Correlation does not prove causation.”

As a consumer of data, it is always important to try to discern if the data is biased and if it seems reasonable. You may not be able to assess accuracy, but you still could possibly identify biases. It is often helpful to ask “Who paid for the study, and why?” 

Mean, mode and median are measures of the central tendency of the data that are typically analyzed. Other areas of analysis such as the measure of standard deviation (the spread of the data around the mean), and outliers are also covered.