We propose two novel diagnostic measures for the detection of influential observations for regression parameters in linear regression. Traditional diagnostic statistics focus on the effect of deletion of data points either on parameter estimates, or on pr