Week 2: research methods ~ statistical thinking

Vocabulary flashcards covering key concepts from the notes on statistical thinking, study planning, data handling, inference, and generalizability.

Statistical investigation

The process of planning, examining data, making inferences, and drawing conclusions to answer a testable question.

Plan the study

The initial step: formulate a testable question and decide who/what/when/how to measure, specifying population, time frame, recruitment, and key variables (and potential confounds).

Population

The entire group of individuals or units to which conclusions will be generalized.

Time frame

The period during which data are collected or the study operates.

Recruitment

The process of selecting or enrolling participants or units into the study.

Key variables

The main factors measured or observed, chosen to address the research question.

Confounds

Variables related to both the exposure and outcome that can bias results if not controlled.

Reliability

The consistency or repeatability of a measurement across repeated trials or raters.

Validity

The degree to which a measurement measures what it is intended to measure (construct validity).

Distributional thinking

Reasoning about the spread and patterns of data across the distribution, not just central tendency.

Confidence intervals

A range around an estimate that expresses uncertainty due to sampling; e.g., 95% CI means the interval would capture the true parameter about 95% of the time in repeated samples.

p-value

The probability, under the null hypothesis, of obtaining results as extreme or more extreme than observed by random chance.

Significance level

The threshold (alpha) used to decide whether to reject the null hypothesis (e.g., 0.05).

Random sampling

A sampling method where every population member has a known non-zero chance of selection, enabling generalization with quantifiable uncertainty.

Generalizability

The extent to which results from the sample apply to the broader population.

Margin of error

The range of error around a sample estimate reflecting sampling variability at a given confidence level (e.g., MOE ±3%).

Random assignment

Assigning participants to groups by chance to balance confounders and enable causal inference.

Causality

Inference that one variable causes changes in another, best supported by randomized experiments.