1
PSY 102: Introduction to Psychology
Lecture 3: Research Methods
Professor Margot Sullivan
https://pollev.com/margotsullivan518 (Poll Everywhere Class Link)
________________________________________
Activity: Why do we need Research Methods? Can you think of some
reasons?
Research methodology is a way to systemically solve or answer a
research problem.
o Selection of the research method is crucial for what conclusions
you can make.
3.1 Guiding Principles for Applying the Scientific
Method to Psychology
Guiding Principle 1
Generalizability- is a measure of how useful the results of a study are
for a broader group of people or situations.
o External Validity.
Need a representative sample of the population.
Population depends on the research question.
“WEIRD” Samples
2010 review of social and behavioral science studies databases
found people from: Western, educated, industrialized, rich, and
democratic (WEIRD) societies represented ~80% of the study
participants.
Random Selection- Procedure that ensures every person in the
population has an equal chance of being chosen to participate.
2
Random Selection Examples
Non-random selection reduces the generalizability of studies
Research Q: How does the average person feel about Drake?
Should we ask 100 randomly sampled people in North America or 100,000
people in Toronto?
Election data: If polling is non-random, election forecasts can be inaccurate.
Guiding Principle 2
Reliability- Consistency of measurement.
A reliable questionnaire should yield similar scores over time = test-retest reliability.
Interrater Reliability
Extent to which different people agree on the characteristics they are measuring.
Example: Two psychologists giving separate diagnoses (low or high IR?).
Example: job interview panel.
Guiding Principle 3
Validity (Construct Validity)- Extent to which a measure assesses
what it claims to measure.
Example: Giving a math test to measure numeracy (low or high validity?).
Example: Questionnaire to measure depression, but it’s really measuring anxiety.
3
Reliability vs. Validity
A test must be reliable to be valid, but a reliable test can still be completely invalid
Example: polygraph
“Distance Index-Middle Width Intelligence Test” Example
Subtracts the width of our index finger from our middle finger.
high test-retest reliability
higher interrater reliability.
BUT, invalid measure of intelligence.
3.2 Types of Research Designs
Descriptive Studies and Correlational Research
Psychologists can’t always exert experimental control over variables.
Example 1: Nutritious vs. non-nutritious maternal diets on the
health of babies.
Example 2: How often do people laugh in the real world?
Only describe patterns of behaviour and discover links/associations
between variables.
(1) Naturalistic Observation
Naturalistic Observation- Watching behaviour in real-world settings
without trying to manipulate the situation.
Conduct in real time or video recordings.
Also, social media.
Wearable technologies.
Robert Provine’s (1996, 2000) Investigation of Laughter Example
Listened to 1200 instances of laughter in social situations.
Recorded:
Gender
Preceding remarks
Reactions
What were the results from the Social Functions of Laughter study?
4
Wansink and van Ittersum (2013)’s Buffet Study Example
Diners could choose between two plate sizes.
Large plate vs. small plate diners
o Chose, consumed, and wasted more food.
Pros and Cons of Naturalistic Observation Studies
High in external validity.
Good starting point.
Can be used to study animal behaviour.
Low internal validity
Cannot draw cause-and-effect inferences
(2) Case Studies
Case Study- Examine one person or a small number of people.
Used to study rare brain damage and/or unusual conditions.
Can inspire new hypotheses about the brain and behaviour.
Henry Molaison (H.M.)
Experimental operation to remove portions of the hippocampus.
Observed by Dr. Brenda Milner from McGill.
Inability to form new memories.
Face Recognition
Prosopagnosia – Extreme impairment of the ability to recognize
faces.
Fusiform face area (FFA) responds specifically to faces.
Pros and Cons of Case Studies
High level of detail of rare psychological phenomenon.
Question: What about external and internal validity?
(3) Self-Report Measures and Surveys
Self-report measures (Questionnaires)- assess participant
characteristics.
o Demographics, personality traits, mental illnesses.
Surveys- Measure opinions and attitudes.
o Market research, political polls.
5
Survey Challenges: People’s responses can vary depending on how the
question is presented.
Open-ended vs. multiple-choice format
Agree/Disagree format (one statement) vs. multiple-choice format
Pros and Cons of Self-Report Measures
Easy to administer.
Get direct information about emotions, personality, opinions.
Potential for dishonesty.
(4) Correlational Designs
Correlation- Strength of association between two variables.
Correlations can be negative, zero, or positive.
Correlation coefficients range from −1.0 to +1.0.
o Higher absolute value means stronger relationship.
6
Correlation ≠ Causation
Illusory Correlations
Perception of a statistical association where none exists.
o Strange behavior and full moon.
o Superstitious behaviour in athletes.
Why We Fall Prey to Illusory Correlations
Humans tend to overemphasize condition A and ignore all other conditions (B
− D).
Bad at remembering non-events.
(5) Experimental Designs
Permit cause and effect inferences.
o Internal validity
Random assignment of participants to conditions.
Manipulation of an independent variable.
Random Assignment
Randomly sort participants into one of two groups.
7
Between-subjects Design
Within-subject Design
Participants are their own control group.
Independent vs. Dependent Variable
IV: Variable that is manipulated.
DV: Variable the experimenter measures to see if the manipulation has an effect.
Confounds
Experiments have high internal validity if the IV is the only difference
between groups.
Confounding variables- Any variable (not the IV) that differs between the
experimental and control groups.
8
Placebo Effect
Improvement because you expect improvement.
Example 1
Can we say that the anxiety drug works in example 1?
Example 2
Participants must be blind to what group they are in.
Example 3
Simply believing nicotine was present, reduced cravings!
Experimenter Expectancy Effect
9
Experimenter expectancy effect- Researchers’ hypotheses lead
them to unintentionally bias the outcome.
Double blind designs (experimenter and participants unaware of
treatment) prevent this.
Demand Characteristics
Cues that participants pick up allowing them to guess the
researcher’s hypothesis.
o “The effects of horror movies on mood.”
o “An investigation of caffeine and memory.”
vs. “Study on University Students’ Film Preferences”
Reactivity
Refers to the tendency of participants to change their behavior when
they know that they are being observed.
Social Desirability Bias/Positive Impression Management
Our tendency to respond in ways that we feel are more appropriate
or socially acceptable to others.
Response sets- Tendency of participants to distort their self-report
responses in a positive way.
Try to impress researcher.
3.3 Statistics in Psychology Research
Psychology and Statistics
Descriptive statistics
Organize data into meaningful patterns and summaries.
1) Central tendency
10
Where the group data tends to cluster, three main measures: mean,
median and mode.
Mean is affected by outliers, but the median and mode are not.
Normal Distribution
2) Variance
How loosely or closely scores are clustered.
Range- difference between the highest and lowest scores.
Standard Deviation- how far each data point is from the mean.
Calculating Range Example:
11
IQ scores Sample 1: 80, 85, 85, 90, 95
Mean = 87; Range: 95 - 80 = 15
IQ scores Sample 2: 25, 65, 70, 125, 150
Mean = 87; Range: 150 - 25 = 125
Range vs. Standard Deviation
Same range, but different standard deviations.
Standard deviation is smaller for a vs. b.
3.4 Research Ethics
Ethical Issues in Research Design
The Tuskegee Syphilis Study (1932 to 1972)
Participants, primarily black men, were diagnosed with syphilis, but not informed.
Not given treatments, even though treatments were available.
Stanley Milgram’s Obedience Experiments (1960s).
Measured the willingness of study participants to obey an authority figure.
Administered “shocks” to a “learner”.
Ethical Guidelines for Research
Research Ethics Boards (REBs) that oversee research on humans and animals.
Informed consent: Participants must know what is involved in the study
prior to participation.
Purpose
Duration
Any potential risks or adverse effects
Right to withdraw
Debriefing
At the end of the study, participants should be fully debriefed about the
true nature of the research.
Should be informed of any deception that was used.
