Shamed for my over three-week delayed reply!

No worries, I just read your reply (remembered this because of discussion in more recent thread, http://www.climateaudit.org/?p=3348 ) .

But if you have any better ideas how to deal with tree-ring based temperature estimate confidence limits, just let us know: we try to recalculate our old data based on your instructions – and also our new one, about 1.5 more replicated data!

I have some ideas, but first I’d like to see a reference where the conventional method is described thoroughly.

Sorry again for my learning to use this forum! My message was supposed to sound something like this:

Shamed for my over three-week delayed reply! But I, unfortinately, have not, during this period, had a possibility to recalculate our seven-year old data. It still seems to me that the confidence limits are most probably properly calculated. We used a simple tree-ring index to temperature conversion to create the modelled average temperature data. Deriving the confidence limits based on this approach, give more accurate results with more replications and vice versa. If you have any better ideas how to deal with tree-ring based temperature estimate confidence limits, just let us know: we try to recalculate our old data based on your instructions – and also our new one, about 1.5 more replicated data! And if you also would like to try with our data, just let us know, and we are ready to consider to cooperate!
Cheers, Mauri!

Shamed for my over three-week delayed reply! But I, unfortinately, have not, during this period, had a possibility to recalculate our seven-year old data. But it still seems to me that the confidence limits are most probably properly calculated. But we used a simple tree-ring index to temperature conversion to create the modelled average temperature data. Deriving the confidence limits based on this approach, give more accurate results with more replications and vice versa. But if you have any better ideas how to deal with tree-ring based temperature estimate confidence limits, just let us know: we try to recalculate our old data based on your instructions – and also our new one, about 1.5 more replicated data! And if you also would like to try with our data, just let us know and we are ready to consider to cooperate! Cheers, Mauri
Cheers, Mauri!

Univariate regression, C of E marriages and alcoholism. Possibly not.

The river waters of England were so polluted for centuries that death was a common result from drinking water. You cannot marry too easily when dead, though some wives accuse husbands of a similar state. The cure was firstly to add alcohol, which killed a number of bugs. Sci-Am once had an article which concluded that for some centuries the usual condition of the European populace was “inebriated”, even down to small children. It had population estimates of alcoholism.

Matters improved a bit when tea came from China before 1800, because boiling the water also killed nasties and tea left a clearer mind than alcohol. Next development was chlorination of public water supplies, which started in London in 1903. We presume that this also caused a lower reliance on grog, but then one has to make a Skill Judgement based on Experience and Undoubted Intelligence and Excellence in Chosen Field of Knowledge.

The question is, with fewer people sozzled, would this lead to more marriages (through people of marriage age being more prevalent and more presentable) or fewer marriages (because sober people might think twice before tying the knot). So, the marriage rate was not a constant – it was perturbed by known secondary effects, calling for multivariate stats.

The interesting question is why the Church of England was selected for the stats comparison, and not another Church. One can predict that the Methodist Church would have a systematic bias, as Methodists did not make love standing up. (Reason? It might encourage them to dance). The Catholic Church was over-represented by people of Irish descent and as a group they drank so much that they would not accurately recall or record if they were married or not, once or more, in sequence or simultaneously.

The closest answer I can give to the Choice of C of E was that they were sexy little devils, with the males well into the art of cherry picking before, during or after marriage.

See, nothing is really very simple in stats.

An old one but a good one:

Storks Deliver Babies (p = 0.008) Robert Matthews
There is a highly statistically significant correlation between stork populations and human birth rates across Europe. While storks may not deliver babies, unthinking interpretation of correlation and p-values can certainly deliver unreliable conclusions.

Reverse the chronological order of the stock prices and try again. It is possible that stock prices have tended to go up recently; see if the reversed order still produces a graph with a recent climb.

Mauri,

About our techniques to calibrate the index to the temperature: Helama, S., Lindholm, M., Timonen, M., Eronen, M. & Meriläinen, J. 2002. Part 2: Interannual-to-centennial variability in summer temperatures in northern Fennoscandia during the last 7500 years extracted from tree-rings of Scots pine. The Holocene 12(6): 681-687

Interesting article, short question: how did you calculate 95 % confidence limits in Figure 1 middle plot? Clearly something different than 2X sample std of calibration residuals.

Here some references you requested:
About our RCS techniques (developed by Samuli Helama):
Helama, S., Timonen, M., Lindholm, M., Meriläinen, J. & Eronen, M. 2005. Extracting long-period climate fluctuations from tree-ring chronologies over timescales of centuries to millennia. International Journal of Climatology 25 (13): 1767-1779.

About our techniques to calibrate the index to the temperature:
Helama, S., Lindholm, M., Timonen, M., Eronen, M. & Meriläinen, J. 2002. Part 2: Interannual-to-centennial variability in summer temperatures in northern Fennoscandia during the last 7500 years extracted from tree-rings of Scots pine. The Holocene 12(6): 681-687

About building the supralong Scots pine chronology:
Eronen, M., Zetterberg, P., Briffa, K., Lindholm, M., Meriläinen, J. & Timonen M. 2002. Part 1: The supra-long Scots pine tree-ring record for northern Finnish Lapland; Chronology construction and initial inferences. The Holocene 12(6): 673-680.

Methodology Page:
We use several statistical approaches in our tree-ring chronology and climate constructions. Because of continuous interest in our Finnish tree-ring studies, I have decided to establish a separate theme page called “Methodology Page” on the Lustia website for describing our way of working with tree-rings. This concerns also the sampling techniques and the sample sizes that play an emphasized role in the case of tree-ring width analyses. Compared to ring width data sampling, isotope and density data sampling, for example, are much more consistent. Making comparisons with different proxies and trying to expose the best proxy variable as terms of the strongest climate signal, is not a very good idea for this reason. Generally, all proxies are important, because they tell about different aspects of climate.