What a statistics major thinks about

Is anything really random?

Most would claim a dice throw yields a random result, as well a coin toss. Classical physicists instead would argue that a given trial of each is in fact not random. The face that shows on the die after the throw is a function the position of the die at that moment, how hard it is thrown, the angle it is thrown, and the characteristics of the surface to which it is thrown on. Likewise, with a coin toss, you can factor in the starting face, the force of the flip (with degree of flipping), and where the coin is caught - high or low. With this information, what the result will be in every trial can be expressed mathematically and correctly calculated each and every time. There is no uncertainty.

I would agree. Uncertainty, after all, is nothing but a deficiency of information. If we knew everything, there would be nothing we did not know. Well, duh.

Inferential statistics exists as a scientific method towards reaching a rational "best guess". We generalize the coin toss and claim that the coin will more or less land on each side roughly half the time. Moreover, we can survey a sample of just several thousand and with high confidence be able to make statements about millions that are very-probably true. The U.S. Census could survey every single person for each of its surveys and then know the exact answer it is attempting to determine, but with what that would cost it would never be able to afford to pay people like me to sit around and do nothing. Also, sometimes it's just not possible to get information about a whole population. This is why they call it a blood sample. You can't take all of it. A statistics professor once gave my class a sampling example using condoms: a certain number of a batch is selected for wear and tear tests. You can't test all of them or else your entire product is ruined! Anyway, it was funny because he didn't seem like a guy to use the word "condom".

Back to the dice example: growing up and playing Monopoly, we at times would try and give the dice a fixed roll (usually just amounting to a drop). We were, though we did not comprehend it at the time, using physics/mathematics/science. By controlling the variables in the "formula" of the dice roll, we hoped to produce the desired result. Sadly, it rarely worked, and also the roll was always followed with, "That was a cheat throw!!!" WWF-style wrestling ensued.

Many are dubious of statistics. Numbers are often flawed by faulty method in obtaining them, and there is always the possibly of error due to freak coincidence. That being said, inferential statistics can give great insight and be very useful. As measurement technology progresses we will have more exact answers available. Yet, until we have complete information on all phenomena, statistical science can aid us in making rationally sound approximations and predictions.