So we can formulate the next hypothesis:
But here we have another problem: How can we operationally define and measure these two variables Solving this problem gives us a prompt to break the dependence between these phenomena, and therefore decrease the level of alcohol consumption and improve academic performance.
Any method of analysis must be based on statistical data. That's why the first stage is gathering any kind of data (statistical, scientific, analytic, forecast) to create an adequate model, which correlates two main variables (alcohol consumption and academic performance) with the help of significant parameters (like student's age, gender, income level, region, alcohol drinks availability etc.).
So, what kind of data can satisfy our goal, can be representative, objective and enough to create an exact, true and accurate model, which can give an opportunity to make a reasonable forecast in future
Of course, such a kind of data should be gathered with the help of questionary. But there is a great variety of questionaries, which differs with size, questions type and canvass methods. Each questionary characteristics set is suitable for certain problem.
In our case we definitely mustn't use a random sample of participants. Why Let's try to explain. ...