Berk: Effect of outliers

In pvrank: Rank Correlations

Description

Data on the average number of births and deaths by the time of the day for a particular hospital in Brussels. The data cover a 30-year period in the nineteenth century.

Usage

data(Berk)

Format

A data frame with 24 observations on the following two variables.

Births

Number of births by hour.

Deaths

Number of deaths by hour.

Details

Twenty-two observations are clustered and show little association. Two observations (for noon and midnight) are dramatically smaller in both the y-direction and x-direction. With these two included, there is obviously a positive correlation in the data. However, a direct association between the two variables is doubtful, for if the outliers are removed then all correlations decrease and the associated p-values increase up to the point where the null hypothesis of independence cannot be rejected at any reasonable level.

Source

Berk, R. A. (1990). "A primer on robust regression". In Fox, J. and Scott Long, J. (Eds), 292–324. Modern Methods of Data Analysis. Sage Publications, Newbury Park, Ca, USA.

Examples

	data(Berk)
	head(Berk)

pvrank documentation built on May 17, 2018, 9:03 a.m.