Audit Data Quality and Data Analytics AAIA Practice Questions (Page 2)

Questions 6–10

1. 6application · medium

   An audit team is assessing the data quality of a dataset used to train a loan approval model. The team wants to identify if there are any unexpected patterns or relationships between variables that could indicate a data entry error or a systemic issue. Which data quality assessment method is MOST suitable for this task?

   • AData cleansing, to correct any identified errors in the dataset.
   • BData profiling, to examine the distribution and relationships between variables like age, income, and loan amount.
   • CPredictive analytics, to forecast future loan default rates.
   • DData quality rules, to check that all loan amounts are within a predefined range.
   Reveal Solution Discussion Select an answer first
2. 7foundation · easy

   An AI auditor completes a data analytics procedure and needs to ensure that the results can be used as evidence to support audit conclusions. What is the most important documentation practice?

   • ADocumenting the process, assumptions, and results
   • BVerbally explaining results to stakeholders
   • CKeeping results in a personal notebook
   • DOnly documenting the final conclusion
   Reveal Solution Discussion Select an answer first
4. 8foundation · easy

   During an AI audit, the auditor discovers that some records in the training dataset contain duplicate entries, while other records have missing values for critical fields. Which data quality dimension is directly compromised by the presence of duplicate entries?

   • AValidity
   • BAccuracy
   • CUniqueness
   • DTimeliness
   Reveal Solution Discussion Select an answer first
5. 9application · medium

   During an audit, the team finds that 5% of the records in a training dataset have missing values for the 'employment_status' field. The team decides to impute the missing values using the mode of the existing values. What is the MOST important step the team must take to ensure the audit trail is complete?

   • ARe-running the model training to see if the imputation changes the results.
   • BEnsuring the imputation is performed using a complex algorithm to improve accuracy.
   • CDocumenting the imputation method, the rationale for choosing it, and the potential impact on the audit conclusions.
   • DDeleting the records with missing values instead of imputing them to simplify the dataset.
   Reveal Solution Discussion Select an answer first
6. 10expert · hard

   An auditor is working with a dataset stored in a SQL database and also has access to the raw log files in CSV format. The auditor needs to perform a complex statistical analysis that requires joining the two datasets and creating new calculated fields. The auditor is proficient in both SQL and Python. Which approach is MOST efficient and maintainable for this task?

   • AUse a data visualization tool to manually combine the data and perform the analysis visually.
   • BUse Python with pandas to load both datasets, perform the join, create the calculated fields, and run the statistical analysis.
   • CUse SQL to perform the entire analysis, including the complex statistical calculations.
   • DUse SQL to join the tables and create the calculated fields, then export the result to a CSV for analysis in a spreadsheet.
   Reveal Solution Discussion Select an answer first