Identifying Misleading Statistics
Sampling Issues
Often times people are mislead because samples are taken that do not represent an entire population. Consider the following example:
In this example we can clearly see that the people being surveyed will not be an accurate representation because those people standing in line early on Black Friday morning will have higher interest levels towards the shopping frenzy than people who are not in line.
Poor Sampling Methods Lead To Increased Bias
A Bit Far Fetched?
The previous example is quite ridiculous but there are many real-life examples of how poor sampling methods lead to bias. In general, if a sample is very convenient (appropriately referred to as a convenience sample) it is usually biased. In addition to this, samples that fail to include major parts of the population lead to bias (this is often referred to as under-coverage bias). Finally, any sample that requires people to respond on there own will lead to bias. People with strong opinions about the topic at hand will be more likely to reply than people who don't care or don't know much about it. Telephone, mail, and radio surveys are all common examples of this.