When data is collected and analyzed we all like to believe that the data is distributed normally which means that there is a particular pattern to the data, however this is not the case in many situations and it is not the end of the world. Non-normal distributions are common and happen much more than some may think. When focusing on data that is normally distributed we immediately gravitate towards methodical tools dealing with statistics such as t-tests, control charts, and the analysis of variance. We look to these graphs and measurements to help us understand our data in a clearer manner. We are often looking for that beautifully shaped bell curve (or something very close to it) that we have all grown to know and love. None the less, if a specific chart or data method is not utilized, normal distribution does not really matter, it only becomes an essential tool requiring normal distribution if statistics are being analyzed. If data is being used in a statistical manner, normal distribution is strongly preferred.
Real Life Situation of Non-Normal Distribution
A real life example of where non-normal distribution might come into place could involve a school setting. Say that a school gets an award for having one of the best science programs around. The school becomes widely recognized as the place to send your children to for an excellent scientific education. However, only a few of the teachers at the school are actually good science teachers exploring the unknown and challenging the students while the rest of the teachers just follow the teachers manual, teach straight from the book and essentially do nothing out of the ordinary that would recognize them as being the best. However, since some of the students tested performed so well and so vigorously on specific science tests, it substantially raised the mean for the whole school, thus making it look as though all students were preforming at an accelerated level when it fact it was really just a few who brought up the whole mean due to their scores. The school and teachers are still rewarded even though it was just a few students who did truly excel beyond the gates of the normal science standards. Non-normal distribution is more common than some may think.
What are the Reasons for Non-Normal Distributions?
There are quite a few different reasons why non-normal distributions many occur. One of the main reasons involves extreme values. Logically, a few extreme values can really offset data, in order to fix this scenario the data must be checked and should be evaluated for things like data-entry errors and measurement errors. If errors are found, those pieces of data should be removed. Another cause of non-normal distribution could include insufficient data discrimination; this means that there are an insufficient number of different values. This is something that can be rectified by either collecting additional data or by implementing more accurate data measurement systems. Furthermore, other contributing factors could include values too close to zero, an overlap of two or more different processes or an ineffective sort of data among other items.