Research Methods and Ethics 02/10/2014
in order for something to be a theory it needs to be falsifiable
what is not empirical research?
You can have a theory but if you test it the wrong way; its not a theory or empirical research
The evidence: if its anecdotal or appeals to tradition
Problem: confirmatory bias, ppl tend to remember when things work out but don’t tend to remember the
times when things don’t work, ex; vitamin C
The Scientific Method
Define the problem
Observer the phenomenon
Form a hypothesis
Test the hypothesis
Primary goal: to establish whether one variable is predicted by one or more other variables. Done with
variables that are internal or essential to the participant, for example gender, sex, etc. (variables that cannot
be changed or are unethical to change)
Ex: does the amount of tv watched a week affect their gpa? (not asking them to watch a certain amount,
just asking them how much they already watch ▯ unethical to ask participants to increase amount of tv
watching etc. bc it could affect their school)
The graph would be a dot graph, xaxis is gpa and yaxis is hours of tv/week
Confounds: uncontrolled additional variable that may correlate with the predictor and affect the outcome.
Predicted variables are sometimes called criterion variables. One confound for example depression (causes
them to watch more tv and not focusing in school which affects gpa). Another variable that is there which
affects both of your variables. May or may not be aware of the confounds.
Big caveat: you can never infer cause and effect. You can say that such and such predicts that this is the
outcome, but you cannot say A causes B, because there could be confounds that you are unaware of. The
result could also have an opposite affect, for example, your marks are low which causes you to watch more
tv because you give up Ecological validity: meaning its easier to study real life situations outside the lab.
Perceived correlations and not actual correlations. We think that a correlation exists based on random
observations whereas it doesn’t.
Crime increases when the moon is full
Stereotypes usually based on this illusory correlations
Primary goal: to establish whether one variable is caused by one or more other variables.
Independent variable: the hypothesized cause. It is the thing that we are manipulating. The researchers
randomly assign participants to the different levels of the variables, while the researchers try to hold
everything else constant. Sometimes the independent variable is equally distributed among participants.
Dependent variable: the thing that we measure and don’t manipulate at all. It is the variable of interest.
Experimental research allows us to infer cause and effect. Rare that you find things that are the sole cause,
so can say that it is one of the causes.
Can sometimes create an artificial situation because it is difficult to take the experiment outside the lab
since you have to manipulate the independent var