Skip to content

CMA Intermediate · Financial Management and Business Data Analytics · Introduction to Data Science for Business Decision-making

An NBFC trains a credit-scoring model on 10,000 past loan applications. Of these, 8,000 were from urban applicants and 2,000 from rural applicants, and historically rural applicants were rejected far more often irrespective of repayment record. The model rejects 10% of urban and 40% of rural applicants. Which statement is correct?

The 30 percentage point rejection gap (40% rural versus 10% urban) points to algorithmic bias inherited from biased historical decisions. The training data and features should be reviewed. Being built on real company data does not make a model fair, and adding more urban data would not cure it.

  1. AThe model is unbiased because it was built on actual company data
  2. BThe 30 percentage point gap suggests possible algorithmic bias inherited from historical data, requiring review of the training data and featuresCorrect
  3. CThe gap of 30% shows rural applicants are inherently riskier, so no action is needed
  4. DThe bias exists only because rural applicants are fewer in number, so adding more urban data will fix it

Explanation

Rejection gap = 40% - 10% = 30 percentage points. Since past rejections were not based on repayment record, the model learns historical bias. Using actual company data does not guarantee fairness, the gap is not evidence of higher risk, and adding more urban data would not remove the bias.

Did you get it right without looking?

One question tells you little. A timed set on Introduction to Data Science for Business Decision-making shows your real accuracy, how long you take and where you lose marks.

Practise Introduction to Data Science for Business Decision-makingFree diagnostic mock

More Introduction to Data Science for Business Decision-making questions