Definitions
Key terms used throughout the Precision documentation. Formal statistical calculations are detailed in the Simple and Complex Statistical Guidelines tabs.
Precision / Imprecision
How closely repeated measurements of the same sample agree. Imprecision is the quantitative expression of that scatter (SD or CV%). Precision studies characterize random error; they do not assess bias or accuracy.
Repeatability (Within-Run Precision)
Variation among replicate measurements obtained under nearly identical conditions within a single run — the baseline analytical noise of the measurement system.
Within-Laboratory Precision
Overall imprecision at a site, combining repeatability with between-run and (when estimated) between-day effects. The most clinically relevant single precision estimate for comparing against a goal or claim.
Between-Run / Between-Day Variance
Additional variability introduced between runs within a day, or across different days of testing. Estimated by Complex Precision long-term (EP05-style) designs via ANOVA.
Simple Precision
A streamlined within-run (short-term) study: multiple replicates in a single run or continuous run, typically at multiple concentration levels, compared against a Goal SD derived from TEa or a manufacturer claim.
Complex Precision
A multi-day study that uses ANOVA to partition variance. Short-term designs (commonly EP15-A3) estimate within-run and within-laboratory precision; long-term designs (commonly EP05-A3) also separate between-run and between-day components.
Observed Mean, SD, CV%
Observed Mean is the arithmetic average of included measurements at a level. Observed SD is the sample standard deviation. Observed CV% = (SD / mean) × 100 — a unitless expression useful for comparing imprecision across levels.
Total Allowable Error (TEa)
The maximum total error (bias + imprecision) considered acceptable for clinical use at a given concentration. Configured in Lab Settings for each assay.
Allowable Imprecision / Goal SD
Allowable Imprecision is the portion of TEa allocated to random error (TEa × Imprecision Allocation%). Goal SD is that allowance converted to an absolute SD at the observed mean, or taken directly from a manufacturer claim when claims comparison is used.
Upper Verification Limit (UVL)
When comparing Complex Precision results to a manufacturer claim, the pass/fail threshold is UVL = F × Claimed CV% (not the raw claim). The F factor accounts for sampling variation in the CV% estimate and reduces false rejection at ~95% confidence.
Levey-Jennings Chart (precision study)
In this module, a plot of standardized results (SDI) in the order they were generated during the study. Used to inspect the collected sequence for shifts, clustering, or extreme points — distinct from Westgard-style multi-rule QC used for ongoing quality control after an assay is in routine use.