Description

This curriculum spans the statistical rigor of a multi-workshop Six Sigma Black Belt program, integrating advanced data analysis techniques with the practical demands of cross-functional process improvement initiatives seen in large-scale operational environments.

Module 1: Defining Project Scope and Aligning with Business Objectives

Selecting critical-to-quality (CTQ) metrics that directly reflect customer requirements and are measurable at scale.
Determining project boundaries by mapping process inputs and outputs using SIPOC (Suppliers, Inputs, Process, Outputs, Customers) under executive constraints.
Negotiating scope with stakeholders when operational processes span multiple departments with conflicting priorities.
Quantifying baseline performance using historical data to justify project initiation and set realistic improvement targets.
Documenting assumptions about data availability and process stability during the Define phase to manage downstream risks.
Aligning project goals with organizational KPIs to ensure leadership support and resource allocation.
Identifying potential regulatory or compliance implications that may restrict process modifications later in DMAIC.

Module 2: Data Collection Planning and Measurement System Validation

Designing a data collection plan that specifies sample frequency, location, and responsible personnel across shifts.
Conducting Gage R&R (Repeatability and Reproducibility) studies for continuous and attribute data to validate measurement reliability.
Selecting between automated data logging and manual entry based on error rates and system integration capabilities.
Addressing missing data protocols by defining imputation rules or exclusion criteria before data gathering begins.
Training data collectors on standardized procedures to reduce operator-induced variation.
Validating time synchronization across data sources when integrating logs from multiple systems or machines.
Documenting calibration schedules for measurement devices to support audit readiness.

Module 3: Process Baseline Performance and Capability Analysis

Testing for process stability using control charts (e.g., I-MR, Xbar-R) before calculating capability indices.
Selecting appropriate capability indices (Cp, Cpk, Pp, Ppk) based on data normality and specification limits.
Transforming non-normal data using Box-Cox or Johnson methods when traditional capability analysis assumptions are violated.
Calculating sigma level from defect rates while accounting for long-term process shifts (1.5 sigma shift convention).
Mapping process yield using first-time yield (FTY) and rolled throughput yield (RTY) across multiple steps.
Identifying outlier subgroups in capability analysis and determining whether to investigate or exclude them.
Reporting baseline performance with confidence intervals to communicate uncertainty in estimates.

Module 4: Root Cause Analysis Using Statistical Tools

Selecting between hypothesis tests (t-tests, ANOVA, chi-square) based on data type and distribution.
Using multi-vari studies to isolate sources of variation across time, location, and part-to-part differences.
Interpreting interaction effects in factorial designs when root causes are not additive.
Validating correlation findings with scatter plots and correlation coefficients before assuming causation.
Applying logistic regression to model discrete outcomes (e.g., pass/fail) against continuous predictors.
Setting alpha levels and power thresholds for hypothesis testing based on risk tolerance and sample constraints.
Handling confounding variables by stratifying data or including covariates in the analysis model.

Module 5: Design of Experiments (DOE) for Process Optimization

Choosing between full factorial, fractional factorial, and response surface designs based on resource limits and factor count.
Randomizing run order in DOE to minimize the impact of lurking time-related variables.
Blocking experimental runs by shift or machine to control for known sources of variation.
Setting factor levels within safe and operable process ranges to avoid safety or quality violations.
Validating model adequacy using residual analysis and lack-of-fit tests after regression fitting.
Optimizing multiple responses using desirability functions when trade-offs exist between goals.
Scaling coded factor coefficients back to real-world units for implementation clarity.

Module 6: Control Plan Development and Sustaining Gains

Selecting control chart types (e.g., p-chart, u-chart, CUSUM) based on data type and sensitivity requirements.
Setting control limits using Phase I data and locking them for Phase II monitoring after process validation.
Defining response plans for out-of-control signals, including escalation paths and corrective actions.
Integrating control charts into existing manufacturing execution systems (MES) for real-time visibility.
Training process owners to interpret control charts and initiate actions without analyst dependency.
Establishing audit schedules to verify control plan adherence during routine operations.
Updating process capability metrics post-improvement to document sustained performance.

Module 7: Statistical Software Implementation and Workflow Integration

Standardizing analysis templates in Minitab or JMP to ensure consistency across project teams.
Automating repetitive analyses using scripting (e.g., Minitab macros, Python with pandas) to reduce manual errors.
Validating software-generated outputs against manual calculations during initial deployment.
Managing version control for analysis files when multiple users contribute to a project.
Configuring software permissions to restrict access to sensitive data or critical templates.
Mapping data pipelines from ERP or SCADA systems to statistical tools using ODBC or API connections.
Documenting analysis workflows to support peer review and regulatory compliance.

Module 8: Change Management and Cross-Functional Communication

Translating statistical findings into operational language for non-technical stakeholders.
Scheduling review meetings with process owners to validate root cause conclusions before implementation.
Addressing resistance to data-driven decisions by co-developing solutions with frontline teams.
Presenting confidence intervals and p-values in context to avoid misinterpretation of statistical significance.
Using control chart dashboards in operational reviews to maintain focus on sustained performance.
Archiving project data and analysis files in a centralized repository with metadata for future reference.
Conducting post-implementation audits to verify that statistical controls remain active and effective.

Module 9: Advanced Topics in Non-Normal and Attribute Data Analysis

Applying non-parametric tests (Mann-Whitney, Kruskal-Wallis) when data fail normality and transformation is ineffective.
Using attribute agreement analysis to assess consistency in subjective inspection processes.
Modeling defect counts with Poisson regression when overdispersion is present in count data.
Calculating process capability for non-normal data using percentile-based methods (e.g., Cpk via Z-scores).
Designing acceptance sampling plans (e.g., ANSI/ASQ Z1.4) with defined AQL and LTPD levels.
Applying time series analysis to detect trends or seasonality in long-term process data.
Validating stability of attribute processes using p- or u-charts with varying subgroup sizes.