Suppose the least-squares regression line for a set of data has slope 72.4. now suppose we remove a point from the data, compute the least-squares regression line, and find that the new slope is 8.7. the point removed would be considered
However, this may or may not be an outlier, depending on the nature of the data. A details on the detection of outliers may be found in an NIST article on outlier detection.
