What to know for research methods?

[sillysmile]

btw, do we need to know about correlation?

Do you mean positive negative correlation?

yeah that's part of it. Although I'm also referring to the strength of the correlation.

[Glockmeister]

I have my own question; could somebody please explain what the hawthorne effect is, and how it differs from other similar extraneous variables. Is it another name for a Self-fulfilling prophecy?

I have to be honest, I'm finding research methods actually quite a bit more complicated then I initially expected.

Hawthorne effect states that a person will act more effectively if they are aware that they are partaking in an experiment. i.e when studying the effect of light on work productivity... if they were told so they'd simply act more productive.

Well, it's not necessarily even told. All you need to induce the Hawthorne Effect is that the people are aware that they are in some sort of experiment or testing salutation. It's the perceived uniqueness and undue attention that increases productivity, results etc...

A Self-fulfilling prophecy is different - it refers to the fact if a person has a belief that a particular result will happen, then the person will have a tendency to unconsciously work towards achieving that result. Where that can become problematic in a result is if the participant has a belief on what the results of the experiment will be.

btw, do we need to know about correlation?

It's not on the current study design (it was in the one beforehand though). Being the last year, I don't think you will get a correlation question.

Nonetheless, it's a rather easy concept to grasp (although it can be really complex, really quick). Basically what a correlation is that there is this relationship between two or more variables. How it works on a conceptual basis is that say a particular person has a certain number of characteristic x (say age) and a certain number of characteristic y (say number of sexual partners per year). What you can do is to plot that data onto a graph (a Cartesian one, that you should be familiar with from doing mathematics). If you have a large enough set of data, what you can is plot all that data together and using either your eyeballs, or some statistical techniques (such as Person's Product-Moment Correlation), you can say whether there is a positive relationship, negative relationship or no relationship between characteristic x and characteristic y (so in our example, there is a positive relationship, negative relationship or no relationship between age and number of sexual partners in a year).

What a positive relationship means is that there is a tendency for those with high of characteristic x to also have high of characteristic y. A negative relationship means that there is a tendency for those with high of characteristic x to have a low of characteristic y. No relationship is that there is no tendency or no link at all between characteristic x and characteristic y. It's probably at good at this point to point out that correlation is obviously not an experimental technique. Consequently, you can not form a cause-effect relationship i.e. Correlation does not imply Causation.