Everyday we, as lay persons, read statistics that we often take for face value. In the newspapers we read in the morning, in the company newsletters we read at work, on the pamphlets we obtain from the doctor s office, on the packages we buy from the grocery store, and during the nightly news we watch in the evening. Many individuals do not stop and think that some of the statistics we read may not be all together accurate because we assume that professionals are providing us with the information and that the research needed to obtain such information has been properly conducted. While this may be true there are other reasons other than inaccurate information that may lead to bad statistics. Sometimes bad reporting of statistics may be deliberate and sometimes it may be a simple matter of error but in any case we should examine the statistics we read a little more closely in order to be sure what we are reading makes sense.
One type of statistics that commonly are plagued with errors are those involving survey results. We often see statistics regarding the results of surveys but never really stop to think about how and who from those results were obtained. In all surveys a certain sample of individuals is provided with questions to answer but what is the demographic of that sample? Is the sample representative of the population as a whole or are there certain groups missing from the sample? If the sample does closely represent the population did all of the individuals targeted for the survey answer? Is the survey biased or does it contain leading questions or questions that an individual may not want to answer correctly? These are all questions that should be considered when considering the results of a sample survey.
Another instance regarding statistics in which we should be more skeptical is the reporting of experiments. Often experiments are conducted in a manner that may produce biased results. For instance, providing a certain treatment for a disease. Researchers may report that a new treatment works but do they also report that it was the only treatment used in the experiment and that other treatments may work just as well? In a clinical trial some groups may refuse to participate more than others and the reporting of an effective treatment may be biased toward one specific group. Finally, experiments with live human subjects often difficult to perform on a moral basis which means in many cases results of experiments may be incomplete due to not being able to perform the necessary steps to get a full explanation.
Finally, another common misreporting of statistics comes from graphs that we see everyday. There are a number of ways in which a graph can be drawn in a way that may be misleading regarding the information provided. In many cases such graphs are in the form of pictograms that can make results look grander than they really are.
Knowing a few ways in which statistic results can be badly reported will help you to analyze a little more closely the statistics you read everyday and will help you to ensure you are reading accurate information.