Statistical Analysis of Laboratory Data

Laboratory managers and supervisors in oil and gas, petrochemical, pharmaceutical, food, water, environmental and manufacturing laboratories

Chemists, analysts and laboratory engineers who generate, review and report test results

Quality assurance and quality control professionals responsible for release decisions and quality systems

Method development and validation specialists who design and document validation studies

Laboratory auditors and technical assessors who review laboratory data and procedures

Classify laboratory data and choose appropriate summary statistics and graphs

Check data sets for errors, missing values and outliers using recognised statistical tests

Select and perform suitable significance tests to compare means, variances and methods

Use regression to build and evaluate calibration curves and assess linearity

Plan and analyse repeatability and reproducibility studies and interpret precision estimates

Evaluate method validation parameters, including accuracy, precision, detection and quantitation limits and range

Estimate measurement uncertainty and apply it when making conformity decisions

Set up and interpret control charts for laboratory quality control and apply run rules

Assess process capability and laboratory performance in proficiency tests

Report statistical results clearly for managers, auditors and customers

Getting the Data Right: Types, Quality and Description

• Why laboratory decisions fail: common data and interpretation errors
• Data types, sampling and the design of data collection
• Descriptive statistics: location, spread and distribution shape
• Graphical tools: histograms, box plots and scatter plots
• Data integrity principles and record control
• Detecting outliers: Grubbs and Dixon tests and median-based methods
• Normality checks and when data need transformation

Comparing Results with Confidence

• Probability distributions used in the laboratory
• Confidence intervals and what they tell decision makers
• Hypothesis testing logic, significance and power
• t-tests, F-tests and paired comparisons of methods and analysts
• Analysis of variance for comparing several groups
• Non-parametric alternatives when assumptions fail

Validating Methods and Measurement Systems

• Fitness for purpose and validation versus verification
• Accuracy, trueness and bias with reference materials
• Precision: repeatability, intermediate precision and reproducibility
• Calibration curves, linearity and regression diagnostics
• Detection and quantitation limits, range and robustness
• Measurement system analysis and gauge repeatability and reproducibility

Measurement Uncertainty and Decision Rules

• Sources of uncertainty in chemical and physical testing
• Bottom-up and top-down approaches to estimating uncertainty
• Using validation and quality control data to estimate uncertainty
• Combined and expanded uncertainty, coverage factors
• Decision rules and statements of conformity with specifications
• Reporting uncertainty to customers and regulators

Monitoring Performance Over Time

• Internal quality control and control charts for laboratories
• Westgard and other run rules: detecting shifts and trends
• Process capability indices and their limits
• Proficiency testing and interlaboratory comparisons: z-scores and follow-up
• Trend analysis, investigations and corrective action
• Automating analysis with LIMS and statistical software, and checking the output
• Action plan: improving data-based decisions in your laboratory

Laboratory data quality and integrity checks

Outlier detection and data screening

Significance testing for method and analyst comparison

Calibration and regression analysis

Precision studies and measurement system analysis

Method validation data evaluation

Measurement uncertainty estimation and decision rules

Laboratory control charts and proficiency test evaluation

Clear statistical reporting of laboratory results