According to the PMBOK® Guide, when dealing with a very large volume of data—such as three million lines of code—it is physically and financially impractical to inspect every single item. In these scenarios, the project manager should use Statistical Sampling.

Efficiency in Large Data Sets: Statistical sampling involves selecting a subset (a " sample " ) of the population of interest (the code) for inspection. The results of this inspection are then used to infer the quality of the entire population.

Reduced Cost and Time: By reviewing a statistically significant sample rather than the full three million lines, the project team can identify systemic issues or high error rates much faster and at a lower cost.

Sample Frequency and Size: The sampling frequency and sizes are determined during the Plan Quality Management process so that the cost of quality (CoQ) is balanced with the level of confidence required in the results.

Why other options are incorrect:

Option A: Pareto chart: A Pareto chart is a histogram used to rank causes of problems from most significant to least significant (the 80/20 rule). While it helps prioritize which errors to fix first, it is not a method for conducting the review or inspection itself.

Option B: Regression analysis: This is an analytical technique used to determine the relationship between variables (e.g., how a change in one area affects another). It is used for forecasting and trend analysis, not for the primary inspection of code quality.

Option C: Automated testing tools: While automated tools are frequently used in software development to run tests, " Automated testing " is not a management tool defined under the standard Quality Management techniques in the PMBOK Guide. Furthermore, code reviews (which check for logic, readability, and standards) often require human or qualitative assessment that simple automated " tests " might miss, making statistical sampling the correct theoretical choice for a management-level inspection strategy.