Essentially, OfCom found that both Lord Monckton and Professor Ball were not sufficiently informed of the nature of the programme, and effectively did not give permission for their interviews to be used; despite this, OfCom found that they were not treated unfairly.

]]>Thanks. That illustrates, though, why comparing trends has little power, as the confidence intervals of the slopes are so large that only a very large difference will be significant.

]]>Bender, we have not yet reached the level of a simple “Er” that does not preface something in the way of clever rejoinder. The Er is a nice way of saying “you simpleton” and the rejoinder explains why.

]]>Just because it’s small doesn’t mean it’s not significant. OLS regression of UAH tropical (land plus ocean) temperature anomalies gives a slope of 0.000410418 degrees/month or 0.049 degrees/decade. The 95% confidence intervals (0.0001356 to 0.000685235 ) do not include zero, the t statistic for the slope is 2.936861204 and the corresponding P value is 0.003527613 so the hypothesis that the slope is not significantly different from zero can be rejected with high confidence. Autocorrelation might expand the confidence intervals and I didn’t check or try to correct for it.

]]>ftp://ftp.ncdc.noaa.gov/pub/data/anomalies/all.ocean.20S.20N.dat

and this data:

http://vortex.nsstc.uah.edu/data/msu/t2lt/uahncdc.lt

indicates that the tropical trend over the oceans is very similar (one would think it would be more) with the satellites showing slightly less (I don’t think it is significant-it is very tiny). ]]>

Has there been any update of the analysis since 1999?

]]>Er …

🙂 ]]>