The least squares line of regression is relatively easy to compute and has a sound theoretical foundation to justify its use, but outliers can have a large effect on the line. The median–median regression line is a more resistant alternative. To obtain the line, divide the data points into three equal groups by size of x (if there is one extra, make the middle group one larger, and if there are two extra make the low and high groups larger). For the low and high groups, obtain the medians of the x and y values separately, giving one summary point for each group. The median–median regression line is the line joining these two points.