Quantitative Data Quality and Levels of Measurement

Vocabulary practice flashcards covering quantitative data quality, psychometric properties (reliability and validity), measurement error, and the four levels of measurement from the NSE 212 Week 6 lecture.

Measurement

The process of assigning numerical values to qualities or attributes using explicit rules.

Psychometric assessment

The formal evaluation of measurement quality, primarily focused on the two core properties of reliability and validity.

Reliability

The extent to which a measurement yields consistent results on repeated measurements under appropriate conditions, associated with consistency, accuracy, and precision.

Test–retest reliability

An assessment of stability concerning whether the same instrument produces stable scores when administered to the same participants more than once under similar conditions.

Internal consistency

An assessment of scale homogeneity that examines whether items within a scale relate to one another and measure the same concept or characteristic.

Cronbachʼs alpha

A numerical reliability coefficient ranging from $0$ to $1$ that describes internal consistency among scale items, with values over $.70$ generally described as acceptable.

Interrater reliability

An assessment of equivalence measuring the degree of agreement between two or more observers using the same tool to record the same behaviour or event.

Intrarater consistency

The degree to which a single observer records comparable behaviour consistently across different occasions.

Validity

The extent to which an instrument measures what it intends to measure, ensuring the interpretation fits the intended attribute.

Content validity

A validity approach that evaluates whether an instrument adequately covers the relevant dimensions of a concept's domain with suitable items and without important omissions.

Face validity

A subtype of content validity based on subjective judgment that an instrument appears to measure the appropriate concept and that its wording is clear and readable.

Criterion-related validity

A validity approach evaluating the relationship between an instrument's scores and a suitable external criterion or reference.

Concurrent validity

A subtype of criterion-related validity that compares an instrument's scores with an external criterion assessed at approximately the same time.

Predictive validity

A subtype of criterion-related validity that evaluates whether an instrument's scores relate to a relevant future outcome.

Construct validity

A validity approach evaluating whether scores behave as theoretical predictions dictate for an intended abstract concept, evidenced by expected relationships, group differences, and scale structure.

Random error

Unpredictable, chance measurement variation (such as transient distraction or an occasional recording mistake) that can obscure patterns in study data.

Systematic error

Constant measurement error that introduces a consistent bias (such as a miscalibrated scale or leading questions), shifting scores or creating misleading differences that cannot be corrected by increasing sample size.

Level of measurement

A classification indicating what assigned numerical values represent and determining which comparisons and statistical operations are justified.

Nominal level of measurement

A measurement level that classifies observations into mutually exclusive categories without rank or numerical amount, summarized using the mode, counts, percentages, and bar charts.

Dichotomous variable

A nominal variable classified into exactly two mutually exclusive categories.

Ordinal level of measurement

A measurement level where categories have a meaningful order or rank, but the distances between categories are unknown and not equal, summarized using the median and mode.

Interval level of measurement

A quantitative measurement level with ordered values and equal numerical distances between units, but lacking an absolute or meaningful zero (e.g., Celsius temperature).

Ratio level of measurement

The highest level of measurement, characterized by ordered values, equal intervals, and a meaningful absolute zero representing the absence of the measured attribute.