Research Methods: Safeguards Against Error

Introduction

Understanding research methods is essential in psychology to ensure that findings are valid, reliable, and free from bias. This section introduces the importance of good research design and the necessity of systematic methodology in psychological research.

The Beauty and Necessity of Good Research Design

Why We Need Research Design

• Research design refers to the structured framework for collecting and analyzing data in a scientific study.

• Good research design helps prevent errors, biases, and misinterpretations in psychological research.

• Example: In the early 1900s, a treatment called facilitated communication was developed for autism. Developers believed autism was a motor disorder, and facilitators would guide a child's hand to type words.

• Initial reports suggested remarkable progress, but later, controlled studies showed that the messages originated from the facilitators, not the children.

• This example highlights the need for rigorous research design to avoid being misled by subjective impressions or wishful thinking.

Additional info: Facilitated communication is now widely discredited, but some still practice it, demonstrating how even intelligent individuals can be fooled without proper research safeguards.

Scientific Methodology: A Toolbox of Skills

Types of Thinking in Research

• Intuitive thinking: Fast, automatic, and based on gut feelings or hunches. Useful for quick decisions but prone to error.

• Analytical thinking: Slow, deliberate, and effortful. Involves careful reasoning and is essential for scientific research.

• Heuristics: Mental shortcuts or rules of thumb that simplify decision-making but can lead to oversimplification and mistakes.

The Scientific Method

• A standardized, systematic approach to studying psychological phenomena.

• Involves forming hypotheses, collecting and analyzing data, and drawing conclusions.

• Helps build and refine scientific theories.

Key Issues in Research Design

Sampling Principles

• Random selection: Every person in the population has an equal chance of being chosen for the study. This increases the generalizability of findings.

• Random assignment: After selection, participants are randomly assigned to different groups (e.g., experimental or control), which helps control for confounding variables.

Reliability and Validity

• Reliability: The consistency of a measurement.

• Types of reliability:

  • Test-retest reliability: Consistency of scores over time.

  • Interrater reliability: Agreement between different observers or raters.

• Validity: The extent to which a measure assesses what it claims to measure. Reliability is necessary but not sufficient for validity.

Replicability and Transparency

• Replicability: The ability to duplicate original findings in new studies.

• Reproducibility: The ability to reanalyze data and obtain the same results.

• Current emphasis in science is on systematic reviews and replication rather than single studies.

Research Designs in Psychology

Naturalistic Observation

• Observing behavior in real-world settings without intervention.

• Advantages: High external validity (generalizability).

• Disadvantages: Low internal validity; cannot infer causation.

Case Studies

• In-depth examination of a single individual or small group.

• Advantages: Provide existence proofs and insights into rare phenomena.

• Disadvantages: Often anecdotal; cannot infer causation.

Self-Report Measures and Surveys

• Questionnaires assessing opinions, attitudes, or characteristics.

• Advantages: Easy to administer; direct assessment.

• Disadvantages: Potential for dishonesty or skewed responses.

• Rating data: Others provide ratings on the target individual, which can help reduce self-report bias.

Correlational Designs

• Examine the extent to which variables are associated.

• Advantages: Can help predict behavior.

• Disadvantages: Cannot infer causation.

• Correlation coefficients range from -1 to 1; the absolute value indicates strength.

• Positive correlation: Both variables increase together. Negative correlation: One increases as the other decreases. Zero correlation: No relationship.

Correlation vs. Causation

• Correlation does not imply causation. Three possible explanations:

  1. A causes B

  2. B causes A

  3. C (a third variable) causes both A and B

Experimental Designs

• Allow researchers to infer causation by manipulating an independent variable and randomly assigning participants to groups.

• Advantages: High internal validity.

• Disadvantages: Sometimes low external validity.

Key Elements of Experiments

• Independent variable (IV): The variable manipulated by the researcher.

• Dependent variable (DV): The variable measured to assess the effect of the IV.

• Confounding variable: Any variable that differs between groups other than the IV and could affect the DV.

Example Table: Types of Variables

Placebo and Nocebo Effects

• Placebo effect: Improvement due to the expectation of improvement.

• Nocebo effect: Harm due to the expectation of harm.

• Blinding participants to group assignment helps control for these effects.

Expectancy Effects and Experimenter Bias

• Expectancy effect: When participants' or researchers' expectations influence the outcome.

• Double-blind designs can help reduce these biases.

Ethical Issues in Psychological Research

Human Research Ethics

• All research must be reviewed by an Institutional Review Board (IRB) to ensure ethical standards.

• Informed consent: Participants must be informed about the study and voluntarily agree to participate.

• Deception must be justified and followed by debriefing.

Animal Research Ethics

• Animal research is used to advance knowledge of brain and behavior, usually with rodents.

• Strict ethical guidelines are in place to minimize harm and ensure humane treatment.

• Only about 7-8% of psychological research involves animals.

Statistics: The Language of Psychological Research

Descriptive Statistics

• Numerical characterizations of data sets.

• Central tendency: Where the group tends to cluster.

  • Mean: Arithmetic average.

  • Median: Middle score.

  • Mode: Most frequent score.

• Variability: How spread out the scores are.

  • Range: Difference between highest and lowest scores.

  • Standard deviation: Average distance of each score from the mean.

Inferential Statistics

• Allow researchers to determine if findings can be generalized to a larger population or occurred by chance.

• Statistical significance: The probability that results are due to chance is less than 5% ().

• Meta-analysis: Combines results from multiple studies to assess the strength of an effect.

• Practical significance: The real-world importance of a finding.

Evaluating Psychological Research

Peer Review and Reporting

• Peer review helps identify and correct flaws in research before publication.

• Common flaws include lack of random assignment, absence of control groups, and experimenter bias.

Media and Scientific Reporting

• Media sources are not always reliable; prioritize information from reputable scientific journals.

• Be aware of sharpening (exaggerating differences) and leveling (minimizing differences) in reporting.

• "Balanced coverage" can sometimes create a false sense of controversy.