...We Welcome You To The Resverlogix HUB withIn The AGORACOM COMMUNITY!

Free
Message: Statistical Significance

Regarding statistical significance...

Definition:

Statistical hypothesis testing is used to determine whether the result of a data set is statistically significant. This test provides a p-value, representing the probability that random chance could explain the result; in general, a p-value of 5% or lower is considered to be statistically significant.

 

Using an practical example...let’s say you wanted to lose weight and so you decided to try some latest and greatest diet.  

 

Before you went on the diet, you decided to weigh yourself first thing every morning for 21 days leading in the diet.  Now, assuming you stayed true to your eating and exercise routine during this pre-diet time, you could still imagine that your weight would change slightly from day to day, even up to a pound or more, but overall would probably still stay within some reasonable min max range.  These small changes or shifts in the data of your daily pre-diet weight is called normal variation and every process has some degree of it.    

 

If you were to graph your pre-diet weight results, it ideally would produce a bell-shaped curve basically meaning that the data is normally distributed meaning some data points are equally above the mean and some equally below with most data points centered near the mean.  

Then comes the diet for however long you go on it after which you once again measure your post diet weight for say another 21 days in the same manner you did pre-diet.  

Now, the reality about data is that there is always some normal variation about it so if you want to conclude that the diet worked meaning that it was statistically significant, then you must basically rule out the possibility of chance, or normal variation.  

 

In other words, in the event that the pre-diet results were biased to the high side due to normal variation and the post-diet results were biased to the low side also due to natural variation, then you could falsely conclude that the diet worked when in fact the results were skewed by the normal variation of the two different data sets.  

 

Therefore, you must statistically calculate and determine if the post-diet results did in fact prove that you lost weight as a result of the diet all other things being the same in your life.  There are statistical methods for determination this that are based on calculating p-values which I won’t elaborate further.

But the bottom line is that when you can conclude that your diet plan resulted in weight loss that was proven to be statically significant, then you have also concluded that the good results must have been achieved by something other than chance meaning that the diet actually did work.

Hope this helps. 

10BagR

Share
New Message
Please login to post a reply