One reason it might be preferable to use the median rather than the mean for simple imputation of numerical data isGroup of answer choicesCalculation of the median is less affected by outliers such as very large values.Calculation of the median is more affected by outliers such as very large valuesThe median is faster to compute than the mean.The median is hardly every used in data science.

One reason it might be preferable to use the median rather than the mean for simple imputation of numerical data is that the calculation of the median is less affected by outliers such as very large values.

Here's why:

1. Outliers: Outliers are extreme values that can skew the data significantly. The mean is sensitive to these outliers. For example, if you have the numbers 1, 2, 3, 4, and 100, the mean would be 22, which is not representative of the majority of the data. However, the median, which is 3 in this case, is not affected by the extreme value of 100.

2. Calculation: When calculating the mean, every value in the dataset contributes equally to the final result. However, when calculating the median, only the middle value (or the average of the two middle values) is considered. This makes the median a better measure of central tendency when dealing with skewed data.

3. Usage in Data Science: While it's true that the mean is often used in data science, the median can also be very useful, especially when dealing with skewed data or when outliers are present. Therefore, the statement that the median is hardly ever used in data science is not accurate.

So, the correct answer is "Calculation of the median is less affected by outliers such as very large values."