A few days ago I was asked about statistical significance and what it really means. This is a concept that often causes confusion for people, so I thought I’d write something for the average person that doesn’t go into too much statistical jargon using words like “rejection region” or “alpha.” Rejection region, by the way, sounds more like the back corner of a discount store where they put the damaged goods on clearance. But it’s a real statistical term and I wouldn't make fun of it because it has a lot of powerful friends who could gang up on me.
So here’s the concept of statistical significance in plain English: It means that there is a consistent difference that is not likely to be caused by accident. It doesn’t mean that the difference is automatically huge or that every observation in one group is different from every observation in the second group. The groups could overlap in content and still be significantly different based on the averages. Statistically significant differences are not necessarily meaningful, either. Think about a five cent difference in cost. This could be statistically significant but still not sway a person to buy the item that is five cents cheaper. So, to sum it up, statistical significance just means that the two groups are consistently different. In other words, the difference (however large or small) is real and probably not due to chance.
Bearing this in mind, I want you all to know that every one of you is statistically significant to me. I check how many people read my blog each week and I know that you are different from other people because you took the time to read this. Knowing that you’ve been here also makes me smile. So this is for you: