Spartan, you’re kind of a remedial aren’t you? Oddly enough you seem determined to display your lack of general knowledge, while using terms which you clearly do not understand. I will try to distil this down and feed it to you in smaller bites. Maybe then you’ll not only learn something but also refrain from displaying your mental ineptitude.
My goal was to show the community two things: 1. how the competitive nature of our program is declining and 2. how the productivity of our program is declining (notably since Zamberlin has been here). Competitive vs. Non-competitive, as I define it, is determined by a 20 Point Margin. I chose this definition because as a team down by 20 you are at the very least 3 touchdowns away from beating the other team. Now that I have this definition I look at the raw data, which doesn’t change, and I count that data to attain the results of "competitiveness". Next Productivity, as I defined it, is the Point Discrepancy between the points ISU scores minus the points ISU allows. The reason I chose point discrepancy is because it eliminates inter-game dynamics (because they are COMMON to both teams) such as weather, location, time of year, injury, ect. Now as a general principle you want to gather as many points of data as you can, this eliminates the effects of
outliers (ex: a fluke 50 point defeat) and allows for
regression to the mean (you should look these up, you might learn something
). In following this principle I added up the entire season’s data on both sides (ISU’s points scored and allowed) and averaged them. In doing this we see regression toward the mean and get a truer picture of what the team looked like across the span of the entire year.
Counting data points does not make my approach flawed, what could arguably make it flawed are the operational definitions. If you want to show that I’m flawed in my research then come up with your own operational definitions, and apply the data you find to those definitions. They can be anything you want!
“sorry - your 'statistical' analysis is flawed if only because you set it up to prove your point
you begin with the premise of 'zamberlin deserves to be fired' and then work to prove the point...
i believe you are committing type 1 error”
Now let’s talk about type 1 error (another comment you made which seems to elude your mental grasp), but first... When beginning any new research one must formulate a hypothesis. Before any research can be done a hypothesis must be established. A hypothesis is the first thing you require before beginning any new research.
My hypothesis was this: Is Zamberlin is a bad coach? This hypothesis can be further defined as: Has ISU become less competitive and less productive under Zamberlin’s supervision?
After my hypothesis was establish I created a null hypothesis, which is contrasted against the alternative hypothesis. What the null hypothesis describes is assumed to be valid unless the actual behaviour of the data contradicts this assumption.
My null hypothesis was this: Is Zamberlin a good coach? Which can be further defined as: has Zamberlin made ISU more competitive and more productive?
This brings me to type 1 and type 2 errors which are used to describe possible errors in the statistical analysis process.
Type 1 error happens if you reject the null hypothesis (Zamerlin is a good coach) when the null hypothesis is true. According to Spartan I committed this error when I: “..begin with the premise of 'zamberlin deserves to be fired' “. This makes little sense as he is now describing the hypothesis process. Furthermore type 1 error can only be committed if I either consciously omit data from the test, ignore the results of the data, or the data represents a statistical anomaly (sometimes referred to as coincidence). That is to say the evidence I presented earlier is only coincidentally displaying less and less “cometitivness” and productivity. This, however, is mathematically unlikely.
Type 2 error happens when you fail to reject the null hypothesis (Zamberlin is a good coach) when it is in fact false. Given the terms of my presentation, I can say that Spartan may be guilty. “Spartan, my boy, you’re in danger of committing type 2 error.”
Admittedly, after the data was gathered and calculated, I made a decision based on my operational definitions of “competitiveness” and productivity. To some this may not be enough to determine one way or another but for me and perhaps others it will be enough to show cause for further investigation (which is the whole point of research).
“statistical evidence tells the strory one way or another - if you go looking for evidence to prove a point one way or another you are likely to find evidence to support your claim no matter what (it is a self fulfilling prophecy).”
A self-fulfilling prophecy (one of the terms you used which seems to be over your head) is a prediction that directly or indirectly causes itself to become true, by the very terms of the prophecy itself, due to positive feedback between belief and behavior. This is a pro-grade (forward moving) phenomenon and for it to make sense in the context you used it I would have to have believe that Zamberlin was already bad for ISU 3 years ago when he got here. Then my belief would have to somehow affect the scores of the games played since he’s been a coach here.
Finally…
“you need to look at the data and see what it says - not decide what you want and set out to prove it”
Once again Spartin you have foot-in-mouth syndrome. I feel that you once watched a 60minutes report on “manipulation of stats in the media” or something and this is now the extent of your knowledge on subjects of anything involving statistical analysis. Here’s the deal, you form a hypothesis, then set out to disprove it while using as much available data as possible. If at the end of your research you are unable to disprove your hypothesis, you must then assume it holds value and credence. In addition research is meant to propagate further research. So if my research is inadequate then you should do further research to shed light on the truth. Just calling me a troll wont be sufficient here.
Here’s what I think Spartin. I think you’re a loyal fan of ISU, but moreover a loyal PAL of Zamberlin, perhaps even super-pals. I know you will find nothing legitimate with my approach, but I challenge you to come up with something of your own.
Come up with something that you can point to that says he is improving ISU’s competitive nature, or productivity, or something completely different of significant importance. Just try it. You might be surprised, and you’ll look like less of an idiot.
