Research Methods: Sampling, Correlational Research, and Simple Experiments

Vocabulary terms and definitions covering Chapters 7 through 10, detailing sampling techniques, correlational research analyses, longitudinal designs, regression, and simple experimental designs.

Population

The bigger, entire set of people from which a sample may be drawn.

Sample

A smaller subset drawn from a population, used to generalize findings back to the population.

Biased Sample

A sample in which not all members of the population have the same chance of inclusion.

Unbiased Sample

A sample in which everybody in the population has an equal chance of being included.

Probability Sampling

A category of sampling involving randomization in some form, used to achieve an unbiased sample.

Non-Probability Sampling

A category of sampling that relies on non-random selection procedures.

Simple Random Sampling

A probability sampling method using an intuitive randomization process, such as a lottery.

Systematic Sampling

A probability sampling method that generally begins with two randomly selected numbers to establish a process of randomization (such as rolling dice).

Cluster Sampling

A probability sampling design where participants are already divided into groups (clusters), and a set of clusters is randomly selected to take part.

Multi-Stage Sampling

A probability sampling method similar to cluster sampling, but where only a portion of each selected cluster is randomly sampled rather than all participants.

Stratified Random Sampling

A probability sampling technique (a random version of quota sampling) where participants are randomly sampled from the population in proportions that approximately match their proportions in the population.

Oversampling

An adjustment in stratified random sampling where a specific group is deliberately overrepresented to enable proper statistical data analysis.

Convenience Sampling

A non-probability sampling technique where researchers collect data from people who are conveniently available to them.

Snowball Sampling

A non-probability sampling method in which existing participants recommend the study to others and spread the word, allowing the sample to grow.

Purposive Sampling

A non-probability sampling technique involving collecting data from a very specific person or demographic in a non-random way.

Quota Sampling

A non-random, stratified sampling method where a researcher recruits a specific target number or percentage of participants per demographic category.

Pearson's $r$ Correlation

A statistical analysis used when both variables are quantitative; the correlation coefficient ($r$) indicates the strength and direction of the association on a scale ranging from $-1$ to $1$.

Chi-Square Test of Independence

A statistical test used when both evaluated variables are categorical.

Spearman Rho Correlation

A correlational analysis designed for ordinal scale data.

Point-Biserial Correlation

A correlational analysis evaluating one quantitative variable and one categorical variable, though often foregone in favor of t-tests or ANOVA.

Phi Correlation

A correlational analysis used when both variables are categorical and dichotomous (two values), though often foregone in favor of the Chi-square test of independence.

Construct Validity

The degree to which the variables in a study are well defined and accurately measure or manipulate the intended concepts.

Statistical Validity

The degree to which the statistical analysis supports the claim being made, taking into account precision, effect size, confidence intervals, and replication.

Spurious Association

An association discovered between two variables that can likely be attributed to a third variable, bringing the meaning of the relationship into question.

Curvilinear Correlation

A non-linear relationship (such as a U-shaped or curved line) that cannot be captured with a straight line and may appear unrelated in a standard linear correlation.

Moderator

An additional variable that impacts the strength or direction of a correlation by specifying conditions, such as who is impacted, when an association applies, or which group shows a stronger association.

Longitudinal Design

A research design that measures variables over multiple intervals of time, turning time into an additional variable without affirming causation.

Cross-Sectional Correlation

A correlation in a longitudinal design testing whether two variables are related to each other at individual, identical points in time.

Autocorrelation

A correlation in a longitudinal design determining whether a single variable correlates with itself over multiple points in time.

Cross-Lag Correlation

A correlation in a longitudinal design testing whether an earlier value of one variable correlates with a later value of another variable, helping establish temporal precedence.

Multiple Regression

A statistical technique evaluating additional predictor variables ($x_1, x_2, x_3$) to help eliminate alternative explanations and rule out third variables when predicting an outcome ($y$).

Parsimony

The scientific preference for the simplest explanation for a phenomenon, requiring fewer assumptions and fewer exceptions to account for.

Mediator

A variable acting as the connecting mechanism or 'glue' that explains why two factors are associated ($x$ leads to the mediator, which leads to $y$).

Independent Variable

The causal variable in an experiment that is deliberately manipulated by the researcher.

Dependent Variable

The outcome variable in an experiment that is measured by the researcher.

Selection Effects

A threat to internal validity in an experiment arising when participants systematically differ between levels of the independent variable.

Matched Groups

An experimental assignment procedure where participants are paired together based on a factor related to the dependent variable, and each member of the pair is assigned to a separate independent variable level.

Independent-Groups Design

An experimental design in which separate, distinct groups of participants are assigned to each level of the independent variable.

Posttest-Only Design

An independent-groups design where participants are randomly assigned to the independent variable and measured on the dependent variable only once.

Pretest/Posttest Design

An experimental design where participants are randomly assigned to independent variable levels and measured twice for each level.

Within-Groups Design

An experimental design where a single group of participants is exposed to all levels of the independent variable and tested repeatedly.

Repeated Measures Design

A within-groups design where participants are measured on the dependent variable after exposure to each independent variable level one at a time.

Concurrent Measures Design

A within-groups design where participants are measured on the dependent variable while receiving each level of the independent variable at roughly the same time.

Counterbalancing

The primary technique used to address order effects by presenting the levels of the independent variable to participants in all possible or random orders.

Demand Characteristics

Cues in an experiment that allow participants to ascertain the study's purpose or hypothesis, potentially altering their behavior.

Manipulation Check

An extra dependent variable incorporated into an experiment to verify that the manipulated independent variable functioned as intended.