Those answers that lead to more questions
Those things that preoccupy, are difficult, frustrating but relevant.Final years as an undergraduate, hmm what fun, statistics, journals, scientific books, data analysis. Don’t get me wrong, I liked science, as a matter in fact, ever since a child, it apealed to me to be wearing those sophisticated white coats . I think it was a cartoon or something that lead me to think it would be cool to be a scientist. That’s right, cool to be scientist (Well I was a kid anyway and coolness was highly subjective of my ego). Anyway I guess in times like this, I’m getting a first hand experience in the scientific arena. And it is bloody frustrating. Yes frustrating, in some cases exciting, and in some other cases darn right boring to the bones.
Whats particualrly interesting and frustrating at the same time is the process. Here we go. Say, I had to learn data analyses. Ok, so I check out some easy to read text books on statistics and data analyses. As I followand travel through the symbols and sentences that resemble statistical logic, a sudden “what the hell is this talking about” pops into mind. I notice some keywords, ok, we have words like “significance”, and ok, so numbers lower than 0,05 or 0,01 is good. It means we have significant relationships, and then you ask your self whats a significant relationship?, and than whats a significant difference. And then you ask your self how do I perform a relationship analysis, and then its like ok, so we have pearson correlation. Ok I like pearson, pearson sounds good, pears tastes good so pearson should be good, so I do a pearson correlation. But then I read a long, watch some SPSS tutorials on youtube and then I ask my self, hmm pearson. Ok than I read the book on statistics and it says “pearson’s coefficient requires paramtric data because it is based upon the average deviation from the mean” and then I think to myself, and then I ask myself, good monkeys, what on earth is parametric data. And good horses, what does deviation from the mean mean? Afterwards I look up parametric data and stuff related to the mean”AHA” “AHA” “AHA”, ok so parametric data are numbers, so say 1 to 2 has a difference of 1 which is the same as 4 to 5 which also has a difference of 1, this refers to parametric data, yes it does, and none parametric refers to ranks, where rank 1 and rank 2 might have differences unequal when comapred to rank 3 and 4 (whatever that means) for an example the value of rank 1 is 10 and value for rank 2 is 8.5, therefore the difference (10-8.5)is 1.5, while the value for rank 3 is 8, and rank 4 is 5, making a difference (8-5) is 3, so the difference between rank 1 and 2 compared to rank 3 and 4 are different . Ok then I recited this a number of times to get it planted in my mind. Ok so I look at my data, and tell myself, Hmm, I think I’ve got some parametric data. And then I read the statistics book again (for the 3rd or fifth time) and watch more youtube vids, interchanging between Frank Zappa and SPSS tutorials. Already hyped up by a condition of endless curiosity and cognitive exhaustion , I came across a subtitle “Assumptions of Parametric Data”. I was like, what the….? Ok so I got some parametric data and I’m already super enthusiastic to play with the pearson, and now this issue of assumptions started ringing. Hmm assumptions, I know assumptions. Then I started to recollect my debating career “Arguments consist of AREL: Assumptions Reasoning Examples and Link back”. Well who knows statistics and debate might converge in harmony. But No, not this time, incorrect access to long term memory I guess. So I read along and it states that all parametric tests have four basic assumptions that must be met for the tests to be accurate. And then I came across new concepts like normally distributed data, homogenity of variance, interval data, independence. And then I said to myslef hmmm I know what normal is, but than after failing to analogize debate and statistics, I reconsidered, oh no, I’m not gona make that mistake again. So I read these four elements and concluded ok, so I need a normally distributed data set, great. Next question. How do I determine whether my data is normal or not? So than I look for another book talking about normality in terms of a Ding DongDing Dong bell curve. And then I came across stuff like transformation, z-score, and more stuff. Most importantly I found the tool to test normality. The name had some russian taste, its called the Kolmogorov-Smirnov test, Hmmm cool sounds like KGB. And so I perform the normality tests. At first I was excited, I recalled my acquaintence with the concept of significance and noted that values below 0.05 or 0.01 are good. So I was like cool, its significant we have a 0.000 which means a good significance level . Than I turned back to my book to find out how to interpret the figures. And then I came across a particularly significant explanation stating that in cases of normality “values with a significance level below 0.05 or 0.01 indicate a significant deviation from normality, indicating that the data is not normal”…..Ummm…. significant deviation from normality??? I had to read this over and over again to combat the prior enthusiasm that had led me to believe that significant levels below 0.05 and 0.01 are always a good thing. And then I finally convinced myself that “young man you have non-normal data!”. “Non-normal data is not good!”. “Non normal data means that you have violated one of the assumptions of parametric data!” Good monkeys…….!!!! This then lead to a series of new questions I had asked myself; what are the consequences of having non-normal data? Could I continue to perform the pearson? When am I got start analyzing something? Is it possible to inhibit my curiosity to stop asking so much questions?
At the time, I could only answer the last question and the answer was “NO”.
I could go on and on and give a dummy’s guide to how annoying statistics are, but I think I’ll end it here. The paragraph above constitutes a microscopic sample of what I have been experiencing lately. Sure, I have gained insights and I shall never regret the hours I have used to read books and discuss these themes with some experts. What I find particularly interesting is my pattern of thought processing. The heart of the question lead to an answer, which will lead to an arteri of questions which will then lead to another answer which will then lead to arterioles of more questions, that will lead to another answer, and that will lead to capillaries of more and more questions. Another thing I find particualrly interesting is the evident conflict between curiosity and brain power. Curiosity seemed to be firing questions every time I had just answered a question. Meanwhile search of an answer had taken up considerable amount of brain power and resulted in brain fatigue. So you eventually have this condition of Mr curiosity (sounds like a song from Jason Mr. AZ) rambling on with questions like What if…?? Whats a…??? How do I…??? Who knows?? And Mr Brain Power begging for mercy “I can’t take it anymore, stop it, stop it, I’m so tired, please understand, I’m out of energy, I need sleep, sleep, sleep, ask me tomorrow” and then Mr Curiosity replies “You still haven’ t answered this?? What about this question?? What are you going to do?? Don’t abandon me, you’ll regret it!!!
Overall science requires enourmous effort and concentration, in addition to food, exercise, sleep and sleeping pills