Skip to content

Lesson 10 — Standards, SPC & Exam Reconstruction

ISO 8000

For Chapter 13, treat ISO 8000 as a supporting concept about application-neutral/portable data meeting stated requirements and the use of standard conventions to define, request, and verify quality. It is lower priority than the core dimensions, rules, lifecycle, and operations.

Statistical Process Control (SPC)

SPC uses repeated process measurements/control charts to distinguish normal/common variation from special-cause change.

A control chart typically includes a center line and upper/lower statistical control limits.

Critical distinction

Statistical control limits are not business acceptance thresholds.

  • Control limits describe expected process variation based on observed behavior.
  • Acceptance thresholds say what the business requires.

A process can be statistically stable and still consistently fail the business requirement.

24 decision rules to reconstruct

  1. fit for purpose vs perfect data;
  2. DQ function vs cleanup project;
  3. criticality vs defect volume;
  4. dimension vs DQ rule;
  5. rule vs metric vs threshold;
  6. validity vs accuracy;
  7. completeness vs integrity;
  8. timeliness vs currency;
  9. consistency vs accuracy;
  10. reasonableness vs validity;
  11. uniqueness vs referential integrity;
  12. profiling vs rule testing;
  13. initial assessment vs monitoring;
  14. profiling vs root-cause analysis;
  15. correction vs prevention vs RCA;
  16. DQ framework vs DQ tool;
  17. prioritization vs improvement goal;
  18. issue assignment vs escalation;
  19. DQ SLA vs scorecard;
  20. parsing vs standardization vs enrichment;
  21. automated vs manually-directed vs manual correction;
  22. controlled manual correction vs direct production patch;
  23. top-down vs bottom-up implementation;
  24. acceptance threshold vs statistical control limit.

Final exam route

  1. Find the threatened business use/consumer requirement.
  2. Ask whether the stem is about criticality, dimension, rule, measurement, issue, or improvement/control.
  3. If a dimension is involved, classify the exact failure before selecting a tool or fix.
  4. Separate baseline assessment from ongoing monitoring.
  5. Separate symptom correction from root-cause prevention.
  6. Identify accountable role/decision mechanism.
  7. Only then choose the tool, technique, or artifact.

Source: pp. 424–470.

← Previous · Exam Map →