Experimental design in statistics
This week I think I will be getting back into Statistics for a while. I am not sure if I truly want to get into this subject, but it would seem that I have at least some interest in it when it comes to playing around with various statistics when it comes to this website. Mainly when it comes to things like traffic, mean word count per post, organic traffic clicks per word, and so forth. However of course there are all kinds of other applications when it comes to statistics, so now and then I do a little more reading on the topic, and work out some code examples when it comes to a few things here and there.
In my travels on line I have come across a Wikipedia article on the topic of something called Experimental Design that struck a nerve when it comes certain things I find myself writing about such as the topic of pure functions. Which prompted me to look into some additional resources on the open web in an effort to gain at least a slightly better understanding of this general topic in statistics.
Whatever I call it something to this effect will come up sooner of later when it comes to starting to play around with a few things when it comes to statistics. One of the first things that is required when it comes to doing something with data is to first have, well, some data. It is best to have some kind of real data to work with, with that said when it comes to my website for example there is Google search console, and Google analytics that help provide some real data when it comes to traffic. In addition I have some of my own scripts that I can use when it comes to tabulating things like word count, mean word count over all, mean word count per category, and so forth.
However what if I want to come up with some kind of hypothesis ( or maybe I should just stick to the word guess sense I am not much of a scientist at this point ) as to the outcome of some kind of action? For example say I take a collection of content that is of very poor quality and invest a solid month of time writing new content while greatly improving the quality of the older content on top of it. Before I make such an investment of time I would like to try to find a way to know if there is a good chance that such an investment of time will end up being worth the effort. In that case I would want to make some kind of projection, based off of some real data, or failing that some kind of educated guess, or even a wild or random guess. This it would seem is where the topic of Experimental Design comes into play. It is a formal way of creating some kind of testable experiment where a research question is asked, and that research question is then tested then an outcome is reached.
However I am pretty new to all of this, so this will be more of an attempt at Experimental Design rather than a serious guide. Which is not always such a bad thing, if an informal approach to something still helps me make smarter choices then mission accomplished.