I am doing an research and trying to choose items for my questionaire ramdomly.But I feel that I am not very clear why I must use ramdomization .
I know that choosing items ramdomly is to prevent the case that there may be some factors contained in a nonrandom selection other than the conditions I set in my experiment would cause the final result ,so that I cann't draw the conclusion that the final result is derived only from the experimental conditions I set.
So I wonder if I can perceive the concept of ramdomization in my experiment in this way : ("-->"means "cause",and "null" means not any particular factor)
(1) null-->selction of items
(2)selection of items , experimetal conditions, some other factors-->final result
By (1)(2),we have:
null,expermental conditions ,some other factors-->final result
Therefore, the selection of the items doesn't account for my final experiment result which is what I intend to make . Is my understanding of random selection in the form:
null-->selection
and how it may affect my experiment result right? I would aprreciate it very much if someone could give me some suggestions !
Your first sentence is I am doing an research and trying to choose items for my questionaire ramdomly. So I assume you are conducting a survey. The aim of the survey will dictate what questions to ask (the data you are collecting). The randomisation enters picture only in deciding whom to ask the questions, or from whom to collect the data (it is not practical to ask everyone).
If your aim is to find out how much a family spends on eating pizzas every month, the question to ask is clear. How much they spend on pizzas and how much is their monthly income (to get an idea of proportion of income spent on this). So your wordings "choose items for my questionnaire randomly" is vague.
Randomisation is needed so that you don't end up collecting the data only from college students or from factory workers, or from professional sports persons(they may be careful about eating out).