JP Morgan Interview Questions | Median or Mode
Question
Give an example where the median is a better measure than the mean
in progress
0
Statistics
55 years
6 Answers
1360 views
Great Grand Master 0
Answers ( 6 )
When a column in the dataset has extreme value ( may be outlires which are important), then median is useful.
For ex: 1 ,1000,5,6 ,2000.
When the data is left skewed or right skewed, then also median is useful than mean.
when the dataset has high degree of skewness then replacement by median is favored
Median and median is largely used in numerical data and more often replacing missing values.
If the numerical column is normally distributed we can replace the missing value with mean of the column
If the numerical distribution is skewed (left/right) a better approach would be replacing the missing values with the median of the column
Suppose, you are trying to have an estimate of salaries earned by a group of 10 people in LPA.
If one of the person earns around 100 LPA and the rest of the 9 people earn around 10-15 LPA.
So, taking the mean will obviously skew the results on the higher side.
It is better to consider median in such scenarios.
When data is normally distributed we can use mean and on other side if data is having high variance or when the data is high skewed(left or right) than we can use median.
When there is a high skewness in the data then it is good to use median. Mean can be use when the data is distributed normally.