Free Isaca AAIA Actual Exam Questions - Question 11 Discussion

A imo, because filling missing data with averages actually changes the original distribution, creating bias that’s hard to detect later. D just loses some detail but doesn’t misrepresent the data as much.

D imo, bins can lose a lot of detail, but A adds artificial data that can skew results more seriously by introducing bias, not just info loss. B and C are mostly cleanup steps, so less risky overall.

A vs D? I’d say D actually causes more info loss because when you bucket numerical data, you’re throwing away exact values and subtle differences. That can lead to oversimplified models and missed patterns. A does risk bias by imputing averages, but it keeps the data continuous and preserves overall trends better than binning does. B and C don’t really mess with the data’s integrity as much, so I’m going with D having the greater combined risk of bias and info loss.

Maybe A has the biggest risk here. When you fill missing values with means or medians, you’re basically making up data that might not reflect actual variance, which can bias the model. Deleting duplicates (B) just cleans data, and splitting attributes (C) usually clarifies things. D does lose detail but binning is often intentional to simplify analysis, not necessarily biased. So, A feels like it could introduce the most hidden errors due to assumptions made on missing data.

I think D has the highest risk since binning can really lose info and skew results more than just filling blanks or removing duplicates.