Psychology Chapter 2: Research Methods

• Chapter 2 discusses the different research methods used in psychology, ethical considerations, and the use of statistics

Good Research Designs

• Highlighted by the example of frontal lobotomies, which lacked systematic research. In 1935, Carlyle Jacobsen presented research related to this topic
• Egaz Moniz developed frontal lobotomy procedures based on this research which were used to treat humans with psychological and behavioral problems
• Moniz was awarded a Nobel Prize in 1949 and Walter Freeman popularized the practice in the USA
• Systematic research on frontal lobotomies effects was not available until the 1960s
• Naive realism and confirmation bias contributed to the practice

Two Modes of Thinking

• System 1 Thinking: It's intuitive, fast, and gut reaction based
• Relies on heuristics (mental shortcuts or rules of thumb)
• Requires little cognitive effort
• System 2 Thinking: It's analytical, slow, and a careful process of evaluation
• Used for reasoning through a problem

Scientific Method

• The scientific method is not a one-size-fits-all process
• It is a toolbox of skills applicable in specific ways based on the context

Random Selection

• Random selection is a technique ensuring everyone in a population has an equal chance to participate in a study
• It increases generalizability of results
• Studying fewer people broadly is more effective than studying more people narrowly

Evaluating Measures

• Reliability is the consistency of measurement using Test-Retest or Interrater Reliability
• Validity is the extent to which a measure assesses what it claims to measure
• Reliability is required for validity, but validity is not required for reliability

Openness in Science

• Open and transparent science ensures findings are replicable and reproducible
• The open science movement was created in response to a replicability crisis in psychology

Replication Crisis

• To resolve the replication crisis, data and materials should be shared publicly
• Conduct replications of original work
• Science journals should publish all sound science, not just flashy studies, using preregistered research

Major Research Methods

• Descriptive
• Correlational
• Experimental

Naturalistic Observation

• Focuses on observing behavior naturally without manipulation and may involve observing subjects like students or animals in their natural habitat
• Major advantages: High external validity to generalize findings and captures natural environments
• Major disadvantages: Low internal validity to draw cause and effect conclusions because of reactivity and observer bias, with no control of variables

Case Studies

• Case studies involve an in-depth analysis of an individual, group, or event
• examples include Patient DF (Mel Goodale, David Milner), Patient HM (Brenda Milner and colleagues), and Patient MC (Striemer et al., 2019)
• Major advantages: Allows investigation of rare phenomena, provides existence proofs, is good for hypothesis generation
• Major disadvantages: cannot definitively determine cause and effect. Generalization and possible observer bias may be an issue

Self-Report Measures

• Research uses interviews, questionnaires, or surveys to gather information related to feelings and attitudes
• Questionnaires measure a variety of characteristics through self-report
• Surveys measure opinions and attitudes through self-report
• Major Advantages: Easy to administer and gather data, cost effective, measures feelings and opinions
• Major Disadvantages: Question writing affects results, accurate insight needed, response sets and malingering alter results

Rating Data

• Rating data are a type of self-report measure where someone else is asked to comment on a person’s behavior, assuming familiarity
• It can avoid malingering and response set bias in self-reporting
• Halo or Horns effects are disadvantages, as well as susceptibility to stereotypes

Correlational Designs

• Research designs measure different variables to see if there is a relationship between them
• They look to see if things co-relate or correlate
• A variable is anything that can vary across or within people such as impulsivity or creativity
• Advantage: flexible, easier to conduct, can determine a relationship between two variables
• Disadvantage: does not explain causation
• It is hard to determine direction of causal relationship or the third-variable problem
• Strength of correlation is measured using a coefficient ranging from -1 to 1. 0 is no relation, and 1 is a perfect relation
• Positive correlations mean variables move in the same direction; negative correlations mean variables move in opposite directions

Correlations and Scatterplots

• Scatterplots show a grouping of data points in 2-dimensions, with each dot representing data from a single participant

Illusory Correlations

• These are variables thought to be related, when they are not, forming the basis of superstitious beliefs
• Can involve lunar-lunacy effects, sports superstitions, and confirmation bias

Experimental Designs

• Defined by random assignment of participants to conditions, and manipulation of at least one independent variable.
• Allows for causal relationship to be established between variables
• Independent variable is manipulated by experimenter
• Dependent variable is measured by the experimenter

Experimental Designs: Key Principles

• Random assignment ensures each participant has an equal chance of being assigned to the experimental/control group
• Between-subjects designs assign different participants to different experimental conditions
• Within-subjects designs have a participant serve as both experimental and control condition
• Treatment A can be a placebo, and Treatment B a new experimental drug

Extraneous Variables

• Researchers try to control for any extraneous/confounding variables such as differences between controls and experimental groups after conditions

Experimental Design Example

• Hypothesis: Drug, "SadBeGone" will increase participants' mood
• Independent variable: "SadBeGone" or placebo
• Dependent variable: Measured mood through a wellness survey

Placebo Effect

• Improvement based on the mere expectation of improvement
• Placebo effects are short-lived
• Control the placebo effect by ensuring the experimental group is blinded to their condition
• Double-blinded experimentation means that neither the researcher nor the participant know which conditions they are in

Nocebo Effect

• Harm from the mere expectation of harm
• Can induce headaches (Morse, 1999) or feeling pain

Experimenter Expectancy

• Hypotheses can unintentionally bias experiments
• Can protect bias through double-blind procedures
• E.g. Clever Hans was a horse who learned to do arithmetic

Demand Characteristics

• When participants try to guess the purpose of a study, they may change their behavior accordingly
• Researchers can disguise the purpose of a study to prevent demand having an impact
• Examples of this involve socially desirable responding

Research Ethics

• Researchers must adhere to strict guidelines
• Ethical questions are difficult, but there are clear cases where past research has not been ethical

Shameful Sciences

• In the Tuskegee Syphilis Study, in 1932-1972 US Public Health Service, ~400 African American men diagnosed with syphilis in Alabama were never offered treatment
• Even after penicillin was proven effective in the 1940s
• During WW2, the Nazis conducted cold water, sterilization, and bone/muscle transplantation experiments on humans

Belmont Report

• The 1979 Belmont Report followed studies in Tuskegee and elsewhere
• Research should
• Allow people to be informed and make decisions
• Maximize benefits and minimize risks
• Distribute benefits and risks to all participants

Research Ethics Boards

• All North American research colleges have at least one research ethics board, REB
• REBs review research and protect participants from harm
• REBs contain members of the institution with expertise in ethics and community members
• In Canada the Tri-Council Policy Statement (TCPS) is followed

Guidelines for Experiments

• Research must provide informed consent, protection from harm, freedom from coercion, a risk-benefit analysis, deception justification, debriefing, and confidentiality

Animal Research

• Institutions engaged in animal research must have an animal research ethics board
• AREBs review programs with animal subjects
• AREBs include faculty, experts, veterinarians, and community members
• The Canadian Council on Animal Care (CCAC) regulates humane use of animals in research

Statistics

• It's an important part of research
• It is the application of mathematics to describe and analyze data
• It is a tool to draw conclusions from data

Descriptive Statistics

• Numerical characteristics of data
• Two main types: Measures of central tendency and measures of variability
• Central tendency: Includes mean (average), median (middle), and mode (most frequent)
• Variability: How loosely or tightly bunched scores are
• Range: Difference between lowest and highest scores
• Standard deviation: how far each data point is from the mean

Descriptive Statistics: Mean, Median and Mode

• Sample IQ scores: 100, 90, 80, 120, 120 (Mean: 102)
• Sample IQ scores: 80, 85, 95, 95, 220 (Mean: 115)

Central Tendency

• The mean tends to provide the best measure of central tendency for normality distributed data sets
• If data is skewed, median/mode works better

Variability: Range and Standard Deviation

• Variability measures how loosely/tightly bunched scores are in the data set: two main methods
  • Range: difference between lowest and highest scores
  • Standard deviation: how far each data point is from the mean (i.e., the average)

Inferential Statistics

• Mathematical methods to generalize findings from a sample population to the general population
• Determine if results are likely to have occurred simply due to chance

Statistical Significance

• The probability that the findings are due to chance
• If the result is statistically significant, it means that the results
• p<.05; or 1 in 20 (i.e., 5%) probability that the observed difference was due to chance alone
• Influenced by sample size, effect size, and variability