PRF flashcards

32 cards. Read the term and the definition here. The deck at the top of this page is optional.

PICO framework

A structured question format specifying Population, Intervention or exposure, Comparator, and Outcome, used to focus a research question and guide the search and design.

Internal versus external validity

Internal validity is the degree to which a study supports causal inference about its own participants; external validity is the degree to which findings generalize to other populations and settings.

Probability versus convenience sampling

Probability sampling gives every population member a known chance of selection, supporting population-level inference; convenience sampling selects accessible participants and supports only limited descriptive inference.

Pilot study purpose

A small preliminary study that assesses feasibility, recruitment, procedures, and variance estimates to refine the protocol, not to test the main hypothesis.

Informed consent elements

Disclosure of purpose, procedures, risks, benefits, alternatives, confidentiality protections, and the voluntary right to withdraw without penalty, followed by documentation of agreement.

Research misconduct types

Fabrication invents data, falsification alters or misrepresents data, and plagiarism appropriates others' words or ideas without attribution; all violate integrity standards.

Vulnerable population safeguards

Groups such as children, prisoners, and cognitively impaired adults require additional protections, including assent, independent monitoring, and justification of their inclusion.

Conflict of interest disclosure

Reporting financial ties, institutional relationships, or intellectual commitments that could bias design, analysis, or reporting, enabling readers to judge potential influence.

Systematic review characteristics

A review using a pre-specified protocol, comprehensive reproducible search, dual independent screening, risk-of-bias assessment, and structured synthesis of eligible studies.

GRADE certainty levels

A framework rating evidence certainty as high, moderate, low, or very low, downgraded for risk of bias, inconsistency, indirectness, imprecision, and publication bias.

Heterogeneity measures

I-squared estimates the proportion of total variability attributable to between-study differences; tau-squared estimates between-study variance in a random-effects meta-analysis.

Publication bias detection

Funnel plot asymmetry, Egger's regression test, and trim-and-fill analysis identify small-study effects suggesting that negative or null findings are underreported.

Reliability versus validity

Reliability is consistency of measurement across occasions, items, or raters; validity is the degree to which an instrument measures the construct it intends to measure.

Cronbach's alpha

An internal consistency coefficient estimating the proportion of total score variance attributable to common variance among items; values above 0.70 are commonly acceptable.

Cohen's kappa

A chance-corrected agreement statistic for two raters on categorical data, computed as observed agreement minus chance agreement divided by one minus chance agreement.

Measurement scale types

Nominal scales label categories, ordinal scales rank, interval scales have equal spacing without a true zero, and ratio scales have equal spacing and a meaningful zero.

Type I and Type II error

Type I error is rejecting a true null hypothesis at rate alpha; Type II error is failing to reject a false null at rate beta, with statistical power equal to one minus beta.

P-value interpretation

The probability of observing data at least as extreme as obtained, assuming the null hypothesis is true; it is not the probability that the null is true.

Confidence interval meaning

A range of plausible values for a parameter; under repeated sampling, 95% intervals would contain the true value in 95% of samples.

Missing data mechanisms

MCAR means missingness is unrelated to any variable; MAR means it depends on observed variables; MNAR means it depends on unobserved values, including the missing value itself.

Multiple imputation

A procedure generating several plausible completed datasets, analyzing each, and pooling estimates with rules that incorporate between-imputation variance; valid under MAR.

Purposive sampling

Deliberate selection of information-rich cases based on defined characteristics relevant to the research question, common in qualitative designs.

Saturation in qualitative research

The point at which additional data collection yields no new codes, themes, or insights; it is specific to the question, population, and analytic approach.

Trustworthiness criteria

Credibility, transferability, dependability, and confirmability are qualitative analogues of internal validity, external validity, reliability, and objectivity.

Mixed-methods design types

Convergent parallel collects quantitative and qualitative data concurrently; explanatory sequential follows quantitative with qualitative; exploratory sequential precedes quantitative work with qualitative inquiry.

Reflexivity

Systematic documentation of the researcher's position, assumptions, and influence on data collection and interpretation, strengthening transparency in qualitative work.

FAIR data principles

Data should be findable, accessible, interoperable, and reusable, typically achieved through persistent identifiers, rich metadata, and clear licensing.

Reproducibility versus replicability

Reproducibility means obtaining consistent results from the same data and code; replicability means obtaining consistent results from a new study with new data.

Codebook components

Documentation of each variable's name, label, data type, permissible values, units, coding scheme, and missing-value conventions.

CONSORT guideline

A reporting checklist for randomized trials covering participant flow, allocation, blinding, outcomes, sample size, and analysis, enabling appraisal and replication.

Number needed to treat

The reciprocal of the absolute risk difference, representing the number of patients who must be treated for one additional beneficial outcome.

ICMJE authorship criteria

Authorship requires substantial contribution, drafting or critical revision, final approval of the version, and accountability for the work's integrity.