Psychology 201W: Introduction to Research Methods
Experimental vs. Non-Experimental Research
- Project Proposals and Ethics Forms are due next week.
- Getting 5 articles to incorporate to your rationale.
- Prior research to support your claims, your hypothesis and your research project.
- Midterm Exam (Multiple Choice: 50-60 Questions) in two weeks. Sign up for
tutorial collection groups during labs that week too.
- Material not covered in lecture but in textbook is your responsibility and will be on
- Lab material is closely related to lecture material.
Review on Last Week’s Lecture:
- Three Approaches to Research: Naturalistic Observation, Experimental and
1. Naturalistic Observation: (Field Studies):
- Decisions to make when running Field Studies: Participant observers
vs. Concealed observers.
- Fundamental Characteristics of Field Studies:
Not testing specific predictions
No attempt to systematically control/alter behaviour
- Does not mean Naturalistic Observation is poor research, it is simply
a different kind of research that yields different types of conclusions.
*Different approaches to research will yield different conclusions and ultimately
different statements that we can make to the relationship of variables.
2. Experiments: The Parable – Hypothetical Example of Coming up
with the Cure for the Common Cold:
- Discussed the number of the issues related to running an experiment
and going through that worked example gives us an idea of how this
process works when we set up an experiment to test the hypothesis
we have to:
Think of all the little problems and in a way debug the
study, and it is a process (doesn’t happen one time).
Have knowledge of research methods – that way the
less likely you are to make mistakes. It is a lengthy
process and they need to have it peer-reviewed by
fellow colleagues to try and identify any potential bugs.
- Have to make some kind of comparison. Experiments involve
comparisons. PSYCH 201W
- Original study of hypothetical research had no control group-
solution: add in a control group
- Expectation Effect of Drug Taking – Solution: “Single Blind” and try
to eliminate the “Placebo” effect.
- Using a placebo to eliminate the confound of drug taking, which can
lead to expectation effect that treatment is going to work.
- Expectation Effect that Treatment is going to Work – to elimination
Experimenter Bias – solution: “Double Blind”
- To make sure that the non-equivalent groups were equivalent –
solution “Random Assignment”
- Characteristics of Random Assignments:
At least two groups that differ systematically on IV
All other variable controlled, to avoid “confounds”
* Bookshelf Analogy: Random Assignment to levels to IV.
You decide how many levels in which you will be manipulating the IV.
Randomly assigning people to different levels of the IV.
E.g. 4 levels of IV (shelves on the bookshelf)
o Participants -> Random Assignment “Books” arranged into the
levels of IV.
o Each of the participants has equal chances of being assigned to
a certain level of the IV.
* Language we use in psychology at times is absolutely of critical importance.
Have to use language properly in order to be consistent in concepts.
3. Non-Experimental Method: Correlational Studies:
- IV is measured not manipulated
- Technique for Controlling Confounds:
Experimental vs. Non-Experimental Research:
1) When differences on Variable A are produced, and then differences on Variable B
are measured. (Experimental Research Approach)
- Experiments are designed to reveal causal relationships. PSYCH 201W
2) When differences on Variable A are observed, and then differences n Variable B
are measured. (Non-Experimental Approach)
- Look to see variables co-vary in some way.
- Designed to reveal predictive relationships.
Class Example: Exam Performance and Anxiety
- Anxiety (IV)
- Exam Performance (DV)
Set up an experiment where all of the students come into the room and are given a
- Except ½ of the students receive in the instructions of the quiz a little note saying:
“This quiz will count for 50% of your grade” (High Anxiety Condition).
- The other ½ of the students receive in the instructions of the quiz a little note
saying: “This quiz is just for practice” (Low Anxiety Condition).
- Take all of the scores and see which of the groups on average perform better and if
there are some results that indicate that the high-anxiety group performed poorly
compared to the low-anxiety group and all other variables are held constant then
we can be somewhat confident in suggesting that anxiety was the cause of the poor
- This would be a relationship and in this case we can say it is a causal relationship
and anxiety caused poor performance.
1/2 Subjects 1/2 Subjects
(Low Anxiety) (High Anxiety)
Average Exam Score Systematically Different; Average Exam Score
Non-Systematic Confounds PSYCH 201W
- Content – Length
Question: Is it ethical to do this?
- No it is not fair or ethical to treat ½ of the students that way and the other ½ of the
students another way to create this high-anxiety.
- You never know how that can affect other people.
- Not really ethical.
This would lead us to study other variables and investigate between these variables
using the non-experimental method of research.
- ½ of the subjects have been randomly assigned to one level of the independent
variable (IV), which is the low anxiety.
- The other ½ of the subjects have been randomly assigned to the other level of the
IV, which is high-anxiety.
- Then we compare the average exam scores of both groups and make sure that
there are no systematic confounds.
Potential Systematic Confounds:
- *Random Error: Some people don’t read the instructions, which could lead to
differences that might seem to reveal causal relationship between those variables,
but in fact there is “random error”.
- Some people are already anxious, but random assignment should take care of that
because everyone has an equal chance of being assigned to one of the two
- What if the two exams are not of equal difficulty level?
If we were to decide that after all it is said and done and the experiment is just
not ethical, we should:
- Think of some other approach to studying this relationship between anxiety and
exam performance, which could be the non-experimental method.
Just before students are about to take the exam, you ask everyone to
fill out a short questionnaire (which measures anxiety).
One measure of anxiety before the students take the exam and then
take the exam and look at the relationship. Is there any kind of
relationship between anxiety levels and performance? Is there any
sort of systematic trend, where the higher the levels of anxiety, the
lower the test scores? PSYCH 201W
Let’s say we ran that study, where we measured anxiety by having students write
the exam and then looked at those variables to see if they co-vary in some way.
- We loo