Master of Wine flashcards

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

Research question

A single, focused, answerable statement of what the study sets out to establish. It should specify the population, the variables, and the scope so that the method and conclusions can be judged against it.

Null hypothesis

A statement that no difference or no relationship exists between the variables tested. Statistical tests assess the probability of observing the data if the null hypothesis were true.

p-value

The probability of obtaining a result at least as extreme as that observed, assuming the null hypothesis is true. It is not the probability that the hypothesis is correct.

Type I error

Rejecting a true null hypothesis, i.e. reporting a difference that does not exist. Its probability is set by the chosen significance level, conventionally 0.05.

Type II error

Failing to reject a false null hypothesis, i.e. missing a real difference. Its probability falls as sample size and statistical power increase.

Statistical power

The probability of detecting an effect of a given size when it truly exists. Power depends on sample size, effect size, variability, and the chosen significance level.

Effect size

A standardised measure of the magnitude of a difference or association, independent of sample size. It indicates practical rather than merely statistical importance.

Confidence interval

A range of values, calculated from sample data, expected to contain the true population parameter at a stated probability, commonly 95 percent.

Sampling frame

The list or register from which sample units are drawn, such as a winery association membership list. Gaps or overlaps in the frame cause coverage error.

Stratified sampling

Dividing the population into mutually exclusive groups, such as region or price tier, then sampling within each group to ensure representation and reduce variance.

Purposive sampling

Deliberately selecting informants who meet defined criteria, such as winemakers with twenty years of experience. It supports depth but limits statistical generalisation.

Convenience sampling

Selecting units that are easiest to access. It is fast and cheap but prone to selection bias, so findings should not be generalised without caution.

Likert scale

An ordinal rating scale, typically five or seven points from strongly disagree to strongly agree. Responses are ordinal, which constrains the choice of statistical tests.

Questionnaire pilot

A small pre-test of the instrument to check comprehension, length, routing, and coding. It reduces measurement error before the main fieldwork begins.

Measurement validity

The degree to which an instrument measures the concept it claims to measure. Face, content, criterion, and construct validity are distinct checks.

Reliability

The consistency of a measurement across items, occasions, or raters. Internal consistency is often reported using Cronbach's alpha.

Triangulation

Using two or more independent methods, data sources, or analysts to examine the same question, so that conclusions rest on convergent rather than single-source evidence.

Literature review

A critical synthesis of prior published work that establishes what is known, identifies gaps, and justifies the new study. It should evaluate, not merely summarise.

Peer review

Independent assessment of a manuscript by specialists before publication. It screens methodology and claims but does not guarantee that findings are correct.

Primary data

Data collected first-hand for the specific study, such as sensory scores or interviews. It is tailored to the question but costly to gather.

Secondary data

Data collected by others for another purpose, such as customs statistics or retail scanner data. Definitions and coverage must be checked for fit.

Descriptive statistics

Summary measures of a dataset including mean, median, standard deviation, range, and frequency distributions. They describe the sample before inferential testing.

Inferential statistics

Methods that use sample data to draw conclusions about a population, including confidence intervals and hypothesis tests such as t-tests and analysis of variance.

Analysis of variance

A test comparing means across three or more groups by partitioning variance. It assumes independent observations, normality, and homogeneity of variance.

Regression analysis

Modelling the relationship between a dependent variable and one or more predictors. Coefficients estimate the change in the outcome per unit change in a predictor.

Correlation versus causation

Association between variables does not establish that one causes the other. Confounding variables and reverse causation must be addressed by design or analysis.

Confounding variable

A third variable associated with both the exposure and the outcome, which can create or mask an apparent relationship if not controlled by design or analysis.

Operational definition

A precise statement of how a variable is measured in the study, such as defining quality by a specified sensory panel score rather than by reputation.

Blind tasting protocol

Serving wines without identifying information so that expectations do not bias judgments. Full blinding also conceals the research hypothesis from assessors.

Triangle test

A discrimination test in which assessors receive three samples, two identical, and identify the odd one. It establishes whether a perceptible difference exists.

Research limitations

Constraints on design, sample, or measurement that bound the conclusions. They should be stated explicitly and their likely direction of bias discussed.

Research ethics approval

Formal review confirming that a study protects participants through informed consent, confidentiality, and secure data handling before fieldwork begins.