CDS Troubleshooting & Problem Resolution 3 — Questions and Answers
Question 1: A data steward is alerted that a nightly data load failed silently — no error was raised but records are missing. What type of control was absent?
- Data encryption control
- Record count reconciliation control (Correct answer)
- Role-based access control
- Data retention policy
Correct answer: Record count reconciliation control
A record count reconciliation control compares expected vs. actual record counts and alerts when they do not match, catching silent failures.
Question 2: Which metric is most appropriate for tracking data completeness issues over time on a stewardship dashboard?
- Average query response time
- Percentage of records with non-null values in mandatory fields (Correct answer)
- Number of ETL jobs run per day
- Data warehouse storage utilization
Correct answer: Percentage of records with non-null values in mandatory fields
Tracking the percentage of records with non-null mandatory fields directly measures completeness, a core data quality dimension.
Question 3: A user claims that a report is pulling 'wrong' data. After investigation, the data steward determines the data is correct but the report filter logic is flawed. This is an example of:
- A data quality defect at the source
- A metadata management failure
- A presentation or reporting layer issue, not a data quality issue (Correct answer)
- A data governance policy violation
Correct answer: A presentation or reporting layer issue, not a data quality issue
When underlying data is accurate but the report produces incorrect results, the defect lies in the reporting logic, not in the data itself.
Question 4: An organization's master data management (MDM) system is generating mismatched golden records. Which troubleshooting step should be taken FIRST?
- Rebuild the entire MDM repository
- Review and validate the matching and survivorship rules configured in the MDM system (Correct answer)
- Disable the MDM system and revert to manual matching
- Increase server resources for the MDM platform
Correct answer: Review and validate the matching and survivorship rules configured in the MDM system
Incorrect golden records are almost always caused by misconfigured matching or survivorship rules, so reviewing those rules is the logical first step.
Question 5: A data steward receives competing requests from two business units, each claiming ownership of the same data element definition. The BEST way to resolve this conflict is to:
- Allow each unit to maintain its own separate definition
- Escalate to the Data Governance Council or steering committee for arbitration (Correct answer)
- Let IT decide which definition to use
- Defer resolution indefinitely until consensus is reached naturally
Correct answer: Escalate to the Data Governance Council or steering committee for arbitration
Data governance councils exist specifically to arbitrate disputes over data definitions, ownership, and standards.
Question 6: During troubleshooting, a data steward finds that a reference table used for code lookups has not been updated in 18 months. This causes downstream analytical errors. The root cause is best categorized as:
- A data security breach
- A master data / reference data management process failure (Correct answer)
- A network connectivity issue
- An end-user training deficiency
Correct answer: A master data / reference data management process failure
Stale reference data indicates a breakdown in the process responsible for maintaining and updating master/reference data.
Question 7: A data quality scorecard reveals that accuracy scores dropped sharply in a particular domain after a system migration. The MOST likely cause is:
- Increased user adoption of the new system
- Data transformation errors or mapping mismatches introduced during migration (Correct answer)
- Improved data entry validation in the new system
- An unrelated change in business KPI definitions
Correct answer: Data transformation errors or mapping mismatches introduced during migration
System migrations frequently introduce accuracy problems due to incorrect field mapping or transformation logic that doesn't preserve the original meaning.
A data steward is alerted that a nightly data load failed silently — no error was raised but records are missing.
What type of control was absent?