PUBH 475 Midterm #2 Study Guide: Research Design and Validity

Cross-sectional data collection

Data is collected once and can only establish an association between variables.

Longitudinal data collection

Data is collected multiple times and can establish causation between variables.

Causation

A change in the independent variable ('cause') directly influences a change in the dependent variable ('effect').

Association

A statistical relationship between two or more variables.

Positive Association

If IV increases, then DV increases, OR if the IV decreases, then the DV decreases (same direction).

Negative Association

If IV increases, then the DV decreases, OR if the IV decreases, then the DV increases (opposite direction).

Data points for causation

Requires two or more data points.

Data points for association

Requires one data point.

Longitudinal study design

Causation is studied using longitudinal designs, involving repeated observations of the same variables over long periods (years or decades).

Cross-sectional study design

Association is examined through cross-sectional designs, analyzing data from a population at a single point in time.

Temporal ambiguity

Arises when it is unclear which variable came first in an observed relationship.

Laboratory research

Takes place in lab settings, easier to manipulate IVs and control experimental conditions.

Field research

Takes place in natural, real-life settings, more challenging to manipulate IVs and control experimental conditions.

Internal validity

The degree of confidence with which we can draw cause-and-effect inferences from a study.

Threats to internal validity

Confounding (extraneous) variables that are not controlled in a study and can potentially influence the results.

History (threat to internal validity)

Events that occur between the pretest and posttest of a research study that could affect participants.

Maturation (threat to internal validity)

Internal influences such as growth changes in mental or physical dependent variables.

Testing (threat to internal validity)

When study participants' survey answers are compromised due to over-testing.

Instrumentation (threat to internal validity)

Measures how well a study can rule out conflicting data results due to researchers changing survey methods.

People (threat to internal validity)

The communication or interaction that occurred between researchers and participants.

Selection (threat to internal validity)

When researchers or participants are given the option to choose the level of involvement in the study.

Mortality/Attrition (threat to internal validity)

Loss of participants due to specific reasons related to the experimental intervention.

Hawthorne effect

The tendency for people to behave differently when they know they are being studied.

Placebo effect

The phenomenon in which the expectations of the participants in a study can influence their behavior.

Diffusion of treatment

Treatment of the experimental group spills over to comparison or control groups.

Location (threat to internal validity)

Differences in the settings or locations where interventions take place, affecting participant responses.

Implementation effect

Differences in persons presenting a program affect the program.

Random assignment

An assignment of study participants to a group where each participant has an equal chance of being assigned to either the treatment or control group.

Experimental study design

Can establish a true cause-and-effect relationship between IV and DV with random assignment.

Quasi-experimental study design

Investigator manipulates the IV(s) or conditions to determine their effect on the DV, but there is no random assignment.

Non-experimental study design

No random assignment, no deliberate manipulation of IV(s), includes most survey research.

Study design nomenclature

X = IV is manipulated; O = observation or measurement of the dependent variable; R = Random assignment.

One-Shot Case Study

Single group observed after treatment without control.

One-Group Pretest-Posttest

Single group measured before and after intervention.

Static Group Comparison

Two groups compared without random assignment.

Posttest-Only Control Group

Control group measured only after treatment.

Pretest-Posttest Control Group

Both groups measured before and after treatment.

Solomon Four-Group

Combines pretest-posttest and posttest-only designs.

Interrupted Time-Series

Repeated measurements before and after intervention.

Nonequivalent Control Group

Comparison of groups without random assignment.

Survey

Systematic investigation to describe group characteristics.

Correlational Research

Explores relationships between two continuous variables.

Causal-Comparative Research

Determines causes of existing differences among groups.

Trend Study

Examines changes in population over time snapshots.

Cohort Study

Same group observed over time for changes.

Panel Study

Same individuals observed over time for changes.

Positive Correlation

Both variables increase together; r closer to 1.

Negative Correlation

One variable increases while the other decreases; r closer to -1.

No Correlation

No relationship between variables; r closest to 0.

Non-Linear Correlation

Curvilinear relationship; r is 0 with U-shape graph.

Causal-Comparative Definition

Also known as ex post facto research.

Formative Evaluation

Assesses program implementation during its conduct.

Summative Evaluation

Assesses final effects or benefits of a program.

Inclusion Criteria

Characteristics required for sample inclusion.

Exclusion Criteria

Characteristics eliminating potential subjects from study.

Sampling Unit

Element considered for selection in sampling.

Sampling Frame

List of individuals from which sample is drawn.

Observation

Process of carefully observing for information.

Sampling Error

Difference between sample statistic and population parameter.

Probability Sampling

Each individual has equal chance of selection.

Nonprobability Sampling

Not everyone has equal chance of selection.

Simple Random Sampling

Every member has equal chance; uses random methods.

Stratified Random Sampling

Population divided into strata for sampling.

Systematic sampling

Selects every nth member after a random start.

Cluster sampling

Divides population into clusters, selects entire clusters.

External validity

Ability to generalize study results beyond sample.

Threats to external validity

Factors limiting generalization of study findings.

Selection threat

Generalization limited by sample's characteristics.

Setting threat

Generalization limited by study's testing environment.

History threat

Generalization limited by specific time period.

Sample size importance

Affects reliability and validity of research results.

Precision approach

Determines closeness of sample statistic to population.

Power analysis

Determines sample size needed for significant results.

Ideal power analysis timing

Before data collection to ensure adequate sample size.

Post-hoc power analysis

Conducted after data collection, less informative.

Hypothesis testing steps

Process includes stating hypotheses and drawing conclusions.

Type I error

Rejecting true null hypothesis.

Type II error

Failing to reject false null hypothesis.

Power

Probability of correctly rejecting a false null hypothesis.

Effect size

Magnitude of experimental effect between variables.

Power and effect size relationship

Larger effect size requires smaller sample size.

Measurement

Assigning numbers or labels based on rules.

Measurement instrument

Tool used to collect data, like questionnaires.

Instrumentation

All measurement instruments used in a study.

Quantitative variables

Measured using numerical values or counts.

Qualitative variables

Measured by describing non-numerical characteristics.

Data collection methods

Self-report, observation, tests, checklists, lab tests.

Self-report strengths

Cost-effective and easy to use.

Self-report weaknesses

Subjectivity and self-awareness limitations.

Self-report biases

Social desirability and recall bias.

Ecological momentary assessment (EMA)

Collects data in real-time within natural settings.

Audio-computer-assisted self-interview (ACASI)

Participants listen to questions and respond digitally.

Questionnaire elements

Instructions, questions, and response options.

Variable characteristics

Includes attributes, values, and relationships.

Levels of measurement

Categorical, ordinal, interval, and ratio scales.

Categorical Data

Data classified into distinct categories.

Nominal Data

Categorical data with no inherent order.

Ordinal Data

Rank-ordered categories without measurable distances.

Continuous Data

Data that can take any value within a range.

Interval Data

Numeric data without a true zero point.