Research and Statistical Analysis

Cognitive Biases and the Post-Truth Era

• Humanity often suffers from cognitive limitations that hinder objective understanding. These include:
    - Hindsight Bias: The tendency to believe, after an event has occurred, that one would have foreseen it.
    - Overconfidence: Excessive confidence in one's own answers and beliefs.
    - Patterns in Randomness: The human tendency to perceive order or meaning in random sequences of data.
• We are currently operating in a "Post-Truth" environment characterized by "Truth Decay," which is fueled by:
    - False news: The dissemination of inaccurate information.
    - Repetition: Hearing a false claim multiple times increases its perceived believability.
    - Powerful examples: Using vivid, anecdotal evidence to overshadow statistical truths.
    - Group identity: Aligning beliefs with a specific social or political group rather than objective facts.

The Foundations of Psychological Science

• To accurately understand and predict phenomena, we must rely on science and research rather than intuition.
• Scientific conclusions are only as valuable as the research methods used to collect the data.
• Critical Thinking: This involves the active exercise of curiosity and skepticism when evaluating the claims of others, as well as one's own assumptions and beliefs.
• Characteristics of Science:
    - Skepticism: A cautious approach to accepting claims without evidence.
    - Empiricism: The practice of relying on observation and experiment.
    - Falsifiability: The requirement that a theory can be proven wrong.
    - Replicability: The ability for a study to be repeated with similar results.
    - Peer review: Evaluation of work by one's peers in the field to ensure quality.
    - Occam’s razor: The principle that the simplest explanation is usually the correct one.

The Scientific Method

• The primary goals of psychology research are to Describe, Explain, Predict, and Change behavior or mental processes.
• Definition: The Scientific Method is a self-correcting process for evaluating ideas with observations and analyses.
• The Cyclic Process:
    1. Theory
    2. Hypothesis
    3. Research
    4. Analysis
• Sources of Error:
    - Bias: Theories can impact observations, leading researchers to see what they expect to see.
    - Need for Precise Operational Definitions: Essential for clarity and measurement.
    - Need for Replication: Necessary for confirmation of findings.

Research Design and Operationalization

• Variable: Any measurable conditions, events, characteristics, or behaviors that are either controlled or observed in a study.
• Operational Definitions: All variables must be clearly defined in measurable terms. For example, instead of defining a behavior as "frequent," an operational definition would specify "$4$ times per hour."
• Hypothesis: A testable statement about the predicted relationship between two or more variables.
• Study Design and Data Collection Questions:
    - What is the best way to answer the research question?
    - How will data be collected?
• Sampling:
    - To be most representative of a population, a sample of participants must be randomly selected from the entire population.
• Analyzing Data and Drawing Conclusions:
    - Data is analyzed using statistics.
    - Hypotheses are never "proved" or "disproved." They are either rejected or not rejected.
    - Conclusions either support or fail to support the hypothesis.

Levels of Analysis and Research Designs

• Research methods vary based on the desired results:
    - Descriptive: Used to describe what occurs.
    - Correlational: Used to test relationships.
    - Experimental: Used to investigate causes.
• Descriptive Designs:
    - Observational: Systematic observation of behavior.
    - Self-Reports: Utilizing various surveys, questionnaires, or interviews to gather information about specific aspects of an individual's experience.
    - Case Studies: An in-depth examination of the experience of a single ($1$) subject.
    - Naturalistic observation: Watching a participant in their environment without interaction.
• Correlational Methods:
    - Used to detect naturally occurring relationships.
    - Assesses the extent to which one variable predicts another.
    - Examines data of $2$ or more variables without intervention or manipulation.
    - Reveals the strength of a relationship.
    - Warnings for Correlational Research:
      - Illusory Correlation: Perceiving a relationship where none exists.
      - Regression Toward the Mean: The tendency for extreme or unusual scores to fall back (regress) toward their average.

Experimental Methods and Variable Types

• Experimental methods examine cause-and-effect relationships and depend on careful design at every step.
• Experiment: A research method in which an investigator manipulates one variable under carefully controlled conditions and observes whether any changes occur in a second variable as a result.
• Variable Types:
    - Independent Variable (IV): The condition or event that an experimenter manipulates to measure its impact on another variable.
    - Dependent Variable (DV): The variable observed to determine the impact of the manipulation of the independent variable.
• Experimental Groups:
    - Experimental group: The subset of individuals for whom the independent variable is manipulated.
    - Control group: The subset of individuals for whom the independent variable is NOT manipulated.

Research Safeguards and Evaluating Findings

• Random Assignment: All members of the sample have an equal opportunity to be assigned to either the control group or the experimental group. This minimizes the placebo effect.
• Double Blind Study: Both the researcher and the participants are unaware of the individual participant's assignment to the control vs. experimental group. This minimizes experimenter expectancy.
• Factors that can invalidate research:
    - Confounds: Variables other than the independent variable that differ between groups and could account for changes in the dependent variable.
    - Placebo effect: Improvement reported because of the expectation of improvement, despite the absence of the independent variable.
    - Experimenter Expectancy: Unintentional bias of the research outcome by the researcher.
    - Demand Characteristics: Participants' guesses about the purpose of the study influence the outcome, leading them to give answers they believe the researcher wants.

Basic Statistics: Descriptive

• Statistics involve using math to describe and analyze data.
• Descriptive Statistics use numerical characterizations to describe information, organize data into meaningful patterns, and reveal what the data "looks like."
• Central Tendency: Provides an index of the most typical score.
    - Mean: The average of all scores.
    - Median: The score that falls exactly in the middle of a distribution.
    - Mode: The most commonly occurring score.
• Measures of Variability: Show how scores vary compared to the most typical score.
    - Range: The difference between the highest score obtained and the lowest score obtained.
    - Standard Deviation ($sd$): The extent to which scores fall away from the mean.
      - A low $sd$ means most scores are near the mean (scores are similar).
      - A high $sd$ means scores are spread out away from the mean (scores are different).

Basic Statistics: Correlation and Inference

• Correlation Coefficients: Measure the nature and strength of the relationship between two variables, or the extent to which one score is associated with another.
    - Scale: Ranges from $-1$ (perfect negative) to $0$ (no correlation) to $+1$ (perfect positive).
    - Direction: Specified by the sign ($+$ or $-$).
      - Positive Correlation ($+$): Both variables change in the same direction.
      - Negative Correlation ($-$): Variables change in opposite directions.
    - Strength: Specified by the number (#). Scores closer to absolute value $1$ are stronger; scores closer to $0$ are weaker.
• Inferential Statistics: Used to test hypotheses and allow inferences to be made about how results apply to larger populations.
• Significance Testing:
    - Null Hypothesis ($H_0$): The position that the two groups will not differ significantly from one another.
    - Fail to reject $H_0$: Occurs when scores are generally about the same (the null hypothesis is likely right).
    - Reject the $H_0$: Occurs when scores are clearly different (the null hypothesis is likely wrong).
• Statistical Significance and Power:
    - Power: The degree of confidence in the conclusion.
    - Significance Level: p < 0.05.
    - Interpretation: There is probably less than a $5\%$ chance the findings are due to coincidence; there is a $95\%$ chance the findings show something real.

The Research Process Overview

• Question: What do you want to know?
• Operational Definitions: What do you REALLY mean by your variables?
• Hypothesis: What do you expect the answer to be?
• Population: Where would this information come from?
• Sample: How do you get info without testing every single person?
• Design: Once you have people, how do you get the info?
• Power: How can you be sure your results aren't due to something else?