![]() In between each point, the data could have been doing anything. They are great at telling a story when you have linear data! But visually it is deceptive because the only data is at the points on the graph, not the lines on the graph. I also used a line graph, which makes the visual connection stand out more than it deserves. Y-axes doesn't start at zero: I truncated the Y-axes of the graphs above.This kind of thing can creep up on you pretty easily when using p-values, which is why it's best to take it as "one of many" inputs that help you assess the results of your analysis. I count each year (minus one) as a "degree of freedom," but this is misleading for continuous variables. When calculating a p-value, you need to assert how many "degrees of freedom" your variable has. To be more specific: p-value tests are probability values, where you are calculating the probability of achieving a result at least as extreme as you found completely by chance. You will calculate a lower chance of "randomly" achieving the result than represents reality. ![]() A naive p-value calculation does not take this into account. If a population of people is continuously doing something every day, there is no reason to think they would suddenly change how they are doing that thing on January 1. Observations not independent: For many variables, sequential years are not independent of each other.Lots of things happen in a year that are not related to each other! Most studies would use something like "one person" in stead of "one year" to be the "thing" studied. This is exacerbated by the fact that I used "Years" as the base variable. If they are related, cool! You found a loophole. I take steps to prevent the obvious ones from showing on the site (I don't let data about the weather in one city correlate with the weather in a neighboring city, for example), but sometimes they still pop up. Noteīecause these pages are automatically generated, it's possible that the two variables you are viewing are in fact causually related. Lack of causal connection: There is probably no direct connection between these variables, despite what the AI says above.It’s a dangerous way to go about analysis, because any sufficiently large dataset will yield strong correlations completely at random. Instead of starting with a hypothesis and testing it, I instead tossed a bunch of data in a blender to see what correlations would shake out. I've been being naughty with data since 2014. That's 636,906,169 correlation calculations! This is called “ data dredging.” Noteįun fact: the chart used on the wikipedia page to demonstrate data dredging is also from me. I compare all these variables against each other to find ones that randomly match up. Data dredging: I have 25,237 variables in my database.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |