January 19th, 2000

 

Research Methods

 

 

Question:  Winning a million dollars would have what effect on your mood?

A.     Make you very happy

B.     Make you very unhappy

C.     Have no effect

D.     Not sure, but would like the chance to find out

 

(Note: many of the definitions below are found in the book)

 

Hypothesis: Winning a million à a good mood.

            We need to define these underlined concepts (million, good mood). Otherwise, we can’t test the hypothesis.

 

Variable Operationalizations:

            “Winning a million”  -- Earning or receiving?

                        Operational definition  -> Publicly winning in a game of skill.

 

            “Good mood” – Self-report vs. behavior?

                        Operational definition à self-report of mood one year later

 

Variable: something that varies in some way and that can be measured or manipulated (changed or altered).

            Height  = variable                     Everybody is a student = not a variable – doesn’t vary because all are students.

 

Dependent Variable (DV): a variable that is merely measured and whose value may therefore DEPEND  on other variables in the study.

 

Independent Variable (IV): a variable that is systematically and randomly manipulated by the researcher. This variable is INDEPENDENT of other variables in the study.

 

Winning a million dollars =  IV

Mood = DV

 

Now, how can we test this hypothesis?

 

Study 1

 

Take all contestants on a TV quiz show and after 1 year, measure their self-report moods.

 

In this case, winning a million is measured (NOT MANIPULATED). (Good) Mood is also measured.

 

There is no IV because there is no manipulation or random assignment to conditions so this is called a:

CORRELATIONAL STUDY.

            Winning a million                       good mood. There is an association but we can’t say that one CAUSES the other.

 

With a correlational study, there are 3 possible types of outcomes:

(1)   positive correlation: as one variable gets larger, the other gets larger. Highest value = 1.0 which represents a perfect positive correlation (seldom observed).

(2)   Negative correlation: as one variable goes up (down), the other goes down (up).  Essentially, the variables have a reciprocal relationship – they move in opposite directions.  –1.0 represents a perfect negative correlation.

(3)   Zero or no correlation: no systematic relationship. It doesn’t appear that the 2 variables move together.

 

If get a positive correlation, can we assume that $  CAUSES (or leads to) Good mood ($ à mood)?

            No!!

Could be:

                        $ à mood

                        $ ß mood

                        $          mood

           

                           being smart :  being smart causes both so it gives the appearance that the two are associated.

 

Essentially, when 2 variables are correlated we don’t know the direction or if one causes the other.

 

 

Study 2

 

Randomly decide who wins 1 million and who wins nothing. Then measure mood 1 year later.

 

Winning million = manipulated = IV

Good mood = measured = DV

 

This is a field study:  manipulated something in the real world.  These are great to do if you can pull it off.

 

Problem:  lawsuit, unethical, difficult to do.

 

 

Study 3

 

Move quiz into lab and randomly assign winners and then measure moods a few months later.

 

Winning $10 = manipulated = IV

Good mood = measured = DV

 

This is a lab experiment and doesn’t’ have the same level of emotional involvement as the field or correlational study (i.e., 10  bucks doesn’t compare to 1 million).

 

Now, can assess if:

$ à good mood

 

There are several possible outcomes to this experimental study (to be found on the web). You should be familiar with what the different outcomes look like and mean.

 

Experimental studies, interpretations:

                           (Assuming a main effect for winning $)

Winning a million à good mod

 

 

Study 4

 

Wouldn't these studies be improved if measured mood before the show and look at mood change.

 

And then, why even look at people who didn’t win?  Just measure change for those who didn’t.

 

This is called:

Pretest-posttest 1-group design

 

Problems:        

(1)   what else happens b/w pre & post-tests?  (e.g., school wins big football game so everybody is in a good mood – you can’t be sure if your results are because of the show or the game).

(2)   Subjects become more experienced at completing the mood measure.

(3)   Subjects may wish to show change or not , depending on their own theories about mood. They may figure out what the purpose of the study is and try to make their results meet what they’d expect.