Question

An investigator wants to test whether exposure to second-hand smoke before 1 year of life is associated with development of childhood asthma (defined as asthma diagnosed before 5 years of age). Give two possible study designs and indicate the pros and cons of each. Then, provide your recommendation for the most efficient design. Which of the following statistics is most resistant to the inclusion of outliers in a dataset?

a) mean
b) median
c) varience
d) standard deviation

Answer 1

b) median

Explanation:

Let's review each measure according to their sensitivity to outliers.

Outliers are values that are very low, or very high.

To not be sensible to outliers, we need a measure that is focused on the centre of a distribution. This measure is the median. In a distribution, 50% of the values will lie below the median, and 50% above. So the median will not have a huge change even if an outlier is included in the data set.

So the correct answer is:

b) median