Earlier, I looked at the HadCRUT3vgl data set using generalized least squares to investigate whether the trend in temperature since 1995 was statistically significant. Here I want to follow-up one of the points from the earlier posting; namely using a statistical technique that fits a local, and not global, model to the...