Lecture Notes on Psychological Research Methods and Theories

A collection of vocabulary flashcards to help review key terms and concepts from the lecture notes on psychological research methods and theories.

Theory

A set of ideas about how the natural world functions, with testable predictions, used to explain empirical observations.

Hypothesis

A testable statement predicting a relationship between variables.

Research method

A method used to test a hypothesis.

Data collection

The process of taking measurements of a method or natural phenomena.

Data analysis

The process of discovering trends and relationships within data.

Report findings

The process of publishing articles in scholarly journals, subject to rigorous peer-review.

Revise theories

The process of integrating new knowledge into scientific understanding, potentially leading to paradigm shifts.

Newton's scientific method framework: Parsimony

The principle of accepting the simplest explanation with the fewest assumptions.

Newton's scientific method framework: Natural order

The principle that the same effects come from the same causes.

Newton's scientific method framework: Generalizability

The principle that mechanisms are universal, applicable from the lab to the real world.

Newton's scientific method framework: Conservatism

The principle that current explanations are accepted until sufficiently challenged by new facts.

Empiricism

The philosophical stance that knowledge requires evidence, not thought alone.

Reductionism

The idea that all mental processes can be explained by neurobiological mechanisms.

Research Design: Within-participants design

A research design where the same participant is tested for both independent variables (IV).

• Cons: Time consuming, high cost.
• Challenge: Practice effect (improved performance by the same participant).
• Benefit: Minimizes confounding variables due to more consistent behavior among the same participant (excluding practice effect).

Research Design: Between-participants design

A research design where the independent variable (IV) is compared across two different groups.

Confounding variable

A variable that obscures the effects of the independent variable, making it difficult to draw conclusions from data.

Reliability

The consistency of input-output in a measurement.

Test-retest method (Reliability)

A method where the same individual performs a test multiple times.

• Considered a poor testing method due to susceptibility to the practice effect.

Interrater reliability

The extent to which multiple human judges assign the same score to subjective criteria.

(Construct) Validity

The relevance of a measurement to the concept it intends to measure.

Operational Definition

How variables are quantified, which is open to argument.

Hypothetical Construct

A theoretical idea that is difficult to measure but has significant implications.

Sampling: Representative sample

A sample of the population that avoids bias and is more realistic (though it may incur higher cost and effort).

Random assignment (Sampling)

A technique to create comparable groups and minimize pre-existing differences, ensuring that results are not due to group composition.

Random sampling

A technique to obtain generalizable results to the overall population.

Participant bias

When a participant's performance aligns with perceived expectations, not necessarily consciously.

Experimenter bias

When an experimenter records performance to align with their hypothesis.

Social Desirability Bias

The tendency of humans to want approval from themselves and others, especially in self-reported data.

Response Set

A tendency to respond in a specific way regardless of the context, often a result of fatigue or experiment length.

Single-blind (Experiment)

An experimental condition where participants do not know which group they are in.

Double-blind (Experiment)

An experimental condition where neither the participants nor the experimenters know the participant's group.

• Benefit: Experimenters cannot unknowingly promote desired outcomes.

Control group

A fundamental component of an experiment, necessary for comparison with the experimental group.

Anecdotal evidence flaws

Limitations of anecdotal evidence:

• A single experience may not be representative of subsequent experiences.
• Personal experience may not be representative of others'.
• Confounding factors are often present and unaccounted for.

Measures of central tendencies

Basic descriptions in statistics that focus on the center of the data, not its distribution.

Mean (Central Tendency)

The average value in a dataset.

• Susceptible to outliers.
• Often the first choice for central tendency, as it allows for more options in inferential statistical analysis.

Median (Central Tendency)

The middle value in a dataset when ordered numerically.

• Robust to outliers (less affected by extreme values).
• Found by rank ordering data (arranging data in numerical order).

Mode (Central Tendency)

The most frequently occurring value in a dataset.

• Typically used for nominal data (categorical data where order doesn't matter) rather than ordinal data.

Standard deviation

The preferred measure of variability (spread) in a dataset.

Descriptive statistics

Statistical methods used to synthesize data into a comprehensible format.

• Assumes acceptable information loss to increase comprehension.
• Based only on the available data.

Histogram

A graphical representation that reports the number of times groups of values fall within a 'bin'.

Bin (Histogram)

A cumulatively defined range on a histogram.

Inferential Statistics

Statistical methods used to infer results about an overall population based on sample results.

Functionally different distributions (Inferential Statistics)

When experimental and control groups are considered different populations (e.g., experimental group influenced by IV, control group representing the general population).

Functionally same distributions (Inferential Statistics)

When experimental and control groups are considered part of the same population (e.g., the general population), implying no real effect of the IV.

T-test

A statistical test that considers all data from two groups to calculate the probability that they come from the same population.

p-value ([0 - 1])

A value produced by a t-test, representing the probability of observing a difference when no real difference exists (i.e., the difference is due to chance).

Statistical significance (p-value < .05)

A universally accepted threshold (p-value less than .05) indicating that an observed difference is likely a true difference, rather than due to randomness or chance.

• Implies less than a $5\%$ probability that there is no real difference and the results occurred by chance.

r correlation coefficient ([-1, +1])

A measure of the strength and direction of a linear relationship between two variables.

• Note: The slope of a regression line is not equal to 'r'.
• Exception: A horizontal line indicates r = 0 (no linear correlation).

Type I error ("False alarm")

The error of observing a difference when no true difference exists.

Type II error ("Miss")

The error of not observing a difference when a true difference does exist.

Observational study

A study that examines variables without manipulation.

• Descriptive and inferential statistics still apply, and it follows the scientific method.

Case study (Observational study)

A detailed examination of a single human.

• Useful for investigating topics with little prior knowledge.

Correlational study (Observational study)

A study with no manipulation, only recording data.

• Hypotheses concern the strength and direction of correlation.

Experiment

A research method designed to manipulate variables and find causal relationships.

Variable

A factor that can take on different values.

• Can be a binary variable (two values) or an operationally defined variable.

Naturalistic observation

An experimental technique where the experimenter observes in the real world, outside of a lab setting.

Confounding variable (Extraneous variable, EV)

A variable that could affect the outcome of an experiment, separate from the independent variable.

• Effects can be minimized by equating groups based on the confounding variable.

Pfungst's falsification of Clever Hans

An example of falsification in science where Pfungst discovered the algorithm behind the horse Clever Hans's apparent intelligence: Clever Hans would lean forward and stomp one more time when the audience became excited, revealing his knowledge of the answer was dependent on audience interaction, not actual mathematical ability.