Identifying outliers in your data. Impact on a Regression Equation

An Outlier is an extreme observation which can be represented by data points that fall beyond the scope of their expected value. Dependent upon the type of data gathered or data being measured, one way to identify "Outliers" (extreme observations) in data would be a Box and Whisker Plot. Although a Box and Whisker plot is not specifically designed to identify outliers, it's a great method to use when shaping large data and/or comparing large groups. Thus, when comparing such large groups with a Box Plot, we employ Whiskers to our boxes that assist in identifying outliers or extreme observations (if present) that fall beyond the scope of the Whiskers or expected values. Outliers impact a regression equation, for example, a linear regression equation, in that they can alter the variation of what is being tested as well as the slope of a regression line.

Businesses/organizations skew the statistics of their research
 
Organizations skew the statistical outcomes of their business research when their research study data and results are exploited merely for their marketing benefit to gain favorable attention or when the method to conduct research is a biased research study to purposely conclude biased results. When the favorable aspect of their research data to increase attention is revealed, the research study is conducted on a specific and biased sample, and the negative aspects or results are not disclosed.

For example, one way that organizations skew statistical outcomes is by abusing the Common Belief Fallacy. Some organizations conduct research studies for their products and/or services on samples of people.  These samples of people are chosen specifically to prove their product or service to be widely accepted. Although these samples of people widely accept their product to be true to its purpose, it does not necessarily conclude or reveal the truth about these products or services.

Secondary Data. A researchers sole source? Why?

First, it could be that some researchers' sole purpose or occupation is to rely on secondary data, concentrate on it, depend on it and analyze by it, meaning that it isn't their occupation to speculate on any other form of data, but the secondary. Some researchers "sole source" is secondary data because the primary research data might have been ineffective, proven ineffective within it's course of research or study, or merely an unreliable source of information, thus requiring other secondary forms of research to bring more reliable, trustworthy and valid results.

For example, with my previous employer, the question "what went wrong the first time, why did the missile hit the water and not the target" under the category of "troubleshooting or control situation (Cooper, D., Schindler, P., 2003, pg. 17)" is "THE" management question that relies on secondary data for more feasible and applicable information.

As a true example, in primary research, our due purpose to shoot a missile is to analyze flight profile telemetry and collect flight profile data during the missile's travel to pursue it's target. Meaning, we shoot a missile so it can tell us how it's operating and what environmental or atmospheric conditions it comes into contact with, during it's flight. With a failed primary attempt to gather "ALL IMPORTANT DATA," we troubleshot the first attempt. However, our troubleshooting was not toward the gathering of any data, but rather to repair an engineering flight issue. So, rather than terminating the whole purpose or program, it would be more cost and research effective to move onto a second attempt to gather such data. In this case and after the second attempt was successful, the appropriate data was gathered and analyzed to be the "sole source" of information, more willingly than the primary.