True experiment: Experimenter manipulates (controls and changes the value of) the independent variable; everything else is held constant.

When we don't manipulate predictor variable, we are doing correlational research

subject variables - non-manipulated variables associated with qualities of the subjects in your study (e.g., IQ nationality, virgin/not virgin).

Which study is most likely to be a true experiment?

a. A study comparing reading comprehension in HIV-positive and HIV-negative 18 year-olds

b. A study that examines who can tell lies more convincingly, 8-year-olds or 13-year-olds.

c. A survey designed to find out if people from the southeastern U.S. are more polite than those from the northeast.

d. A test of performance to see if athletes perform better when their coaches are present or when their coaches are not present.

If we don't manipulate our predictor variable, we can't say for sure whether or not it CAUSES our outcome variable.

Correlational Research - research that examines the degree to which two variables are related, so that knowing the value of one allows us to predict the other

- but we don't know whether one causes the other.

correlation is not equal to causality!!!

If 2 variables are correlated, when one changes, the other does too.

Correlation coefficient - measure of how closely the values of two variables are related to each other.

ranges from -1.0 to 1.0

positive correlation - when one number goes up, the other goes up

negative correlation - when one number goes up, the other goes down

perfect correlation - one variable can be EXACTLY predicted from the other (e.g., -1.0 or 1.0 correlation)

possible causal patterns when 2 variables are correlated:

a causes b

b causes a

c causes a & b

Practice Question!

As the number of churches goes up in a town, so does the number of crimes committed. From this relationship, you can definitely conclude:

a. There is a negative correlation between number of churches and number of crimes.

b. Building more churches causes a higher crime rate.

c. An increase in population causes more crime and produces a greater need for churches

d. None of the above

Causation implies correlation:

If something CAUSES something else, it should be correlated with it

Ex. If dancing causes me to sweat, then whenever I dance, I should sweat

BUT correlation doesn't imply causation

Confound or Confounding Variable - variable that produces an effect that is confused (confounded) with the effects of the independent or predictor variable

Confounds can occur when we don't manipulate Indep. variable or don't hold everything else constant

Placebo Condition - Experimental condition that research participants think or expect will have some effect, but which in fact is inert (has no effects).

operationalize - turn an abstract concept into a variable that can be measured or manipulated

Problems with self report:

- only good if report is honest

- social desirability bias

- some people can't report

behavioral measures

- does it really capture concept?

rating

- maybe too subjective

- would other people agree?

reliability: stability or consistency in measurement

inter-rater (or, inter-observer) reliability -

Degree to which two (or more) raters or observers agree that they have seen the same thing.