Amaleq13

Frankly, I'm still somewhat confused about the particulars of the disagreement since the article doesn't mention the "sigmas".


I think you may be the only member here who can refer to me as living "down south". It hasn't started snowing yet but the termination dust is halfway down the mountains around my house so it won't be long.

andrewcriddle

IIUC the one sigma (one standard deviation) range is the 68% confidence limit and the two sigma (two standard deviations) range is the 95% confidence limit.

Andrew Criddle

spin

That is correct, Andrew.

This is why one sticks to citing the 2 sigma ranges.

ETA: And I note Doudna's BibInterp article is dated 2004. He seems ignorant of John Deere's unsupported date-range for pHab. Perhaps John Deere invented it and that is the reason he will never site his data source for the information, but instead accuses me of misinformation in his "the ludicrous conclusion". This doesn't augur well for the reliability of anything he writes if he is so easily deceived by anonymous tipsters.


spin

rlogan

That much is true, Andrew.

I am trying to see how people are deriving these. They are not symmetric, as can be seen by simply looking at the ranges. So they are not based on t-distributions or normal distributions.

A 95% confidence interval from, say, a normal distribution is

point estimate + or - (1.96*sigma)

That is why they are "eyeballing" it and calling it "two sigma"



For my buddy down there in the deep south:


Any estimate varies from sample to sample. One might say 95BCE and another might say 30BCE, and another 60BCE.

A way of measuring this variation is denoted "standard deviation", or "sigma".


For two very common sampling distributions, about 68% of the observations will be within one standard deviation of the true mean. 95% of the observations will be within two standard deviations of the true mean.

If you came up with a point estimate of, say, 50 BCE and a standard deviation of 25, then your 95% confidence interval for the actual date would be 50 + or - 2*25

That is, from 100 BCE to 0CE.


We can use this information to formally test hypotheses about proposed dates.

In this example, any proposed date within the range 100BCE to 0CE cannot be rejected at the 95% level of confidence (5% critical level).

On the other hand, we can reject a proposed date of 30CE or a proposed date of 150BCE at that level of confidence.



Again, I can't see the formal statistics backing these dates. I have a paper I have been sent, for example, where the dates are just thrown out there and I cannot see how they are derived.

I know how to do these, if I am given the underlying distribution and sampling data. Since the intervals I see in these papers are not symmetric I am not sure whether these are based on noncentral distributions or ones with some tricky skewness and kurtosis or what. I think its some kind of non-continuous thing from the calibrations. But I need to see it.

John Deere

Toto:

In the article - which is not from the source I was told about - Greg appears to simply be using the "prerecalibrated" results.

He wrote:

"In radiocarbon datings of Dead Sea texts at Zurich in 1991 and Tucson in 1994,"

As there are no data sets I can only assume he was simply refering to the original results. However this may be, something seems screwy with his data as in the following quote he seems to give the correct "recalibrated" 68% confidence sigma 88-2BCE - the one we used in the article - but the incorrect "unrecalibrated" 95% sigma 160BCE -2BCE - notice the 2 sigma range extends 72 years to the older date but only 4 years toward an earlier one.

"1QpHab, which is almost certainly contemporary to 4QpPsa, radiocarbon dated earlier, in the 1st century BCE (160 BCE-2 CE at 95% confidence [at 68% confidence, 88-2 BCE])."

I am curious as to how Greg's math works and am attempting to reach him and get his data. I will post my findings.


Joe




Joe

spin

Still no revealing of how he obtained his date range for Pesher Habakkuk.

John Deere

Spin:

Allow me to apologize I was flat wrong.

A friend has just read me over the phone Greg's article from DDS After Fifty and the two sigma range he gives has its younger edge as 2CE, just as you said. He says he can provide the related data, so I will be able to compare how we diverged.

Joe

rlogan

That is what I was saying about the nonsymmetry. When you reach him I would like to know why.

John Deere

I have finally obtained Doudna’s C14 results (‘Table A’ in his piece in The DDS After Fifty Years) and wish to make the following comments.

In the article I wrote C14 results were given for the following Scrolls: 4Q266, 1QpHab, 1QS, 4Q258, 4Q521, 4Q267, 4Q208, 4Q22, 4Q2, 1QIsa, 4Q171, Kefar Bebayou, and Hev 21. These Scrolls were also given C14 results in Doudna’s above mentioned data set. This created 52 shared data points, in other words each Scroll had four numbers attached to it, these being the youngest and oldest one sigma dates and the youngest and oldest two sigma dates, thus 13 times 4 or 52. Of these fifty two shared data points 51 are within one percent of one another and one is within two percent. This being the forty years difference between the youngest edge of the two sigma range in 1QpHah, a point I shall return to shortly.

Readers of this thread know that I have been accused of fabricating my data, with the clear implication that Doudna’s data was reliable and mine was not. If the above perfect overlay was the result of fabrication, all I can say is that my imagination must be pretty sensational.

I wish to make the following request to the moderator. That he or she review the overlay between the two data sets, confirm my findings and following that delete every posting by Spin attacking my character. If this is not done I will never place another post on this thread. Spin, obviously, hurled his accusations, not from any nuanced comparison of the two data sets, which when compared as a group are the same, but merely from his observation of a single discrepancy. Such an approach is simply not rational and since you deleted my comment of Spin’s mental state I would ask that you now do the same to his discredited postings.

Returning to 1QpHab I would first point out that only one of its four data points is out of alignment and this by 2% - 40 years being 2% of 2000. There are two possible causes for this discrepancy. Doudna’s used a different curve, (Unlikely given the perfect overlap in the other dates) In both our articles Doudna and I point out that a number of different ‘curves’ may be used and that they could produce results as different as the one seen here. The other possible cause would be an error in transcription or arithmetic. (Likely) This is probably the source as the two numbers in question are 4 and 40 and errors involving a trailing 0 are common.

If the error is mine I shall place an errata notice to Dead Sea Discoveries disclosing it. Further, if it can even be shown that Doudna used a different curve that produced a different result but is nevertheless a valid one – as seems obvious from the overlays – I will write them a letter mentioning this, as it might be of interest to readers of my article. Having said this I would also point out that in either case I will not need to change a single word of the article. The error – if it exists – is not mentioned in the text and appears only as a ‘tail’ in the graphic representation of 1qphab’s two sigma range. From the perspective of our thesis it makes absolutely no difference which two sigma range for 1qpHab I use, Doudna’s or the one given in our article. The argument in the article stands in either case, every word is correct, and every source is accurate.

This is made completely clear by the following comment by Doudna regarding 4QpPs. That Scroll presents a far worse problem for his thesis – that all of the Scrolls were written mid-first century BCE - than 1QpHab does for our argument, which merely posits that C14 dating has not ruled out that the sectarian Scrolls could have been written in the first century. It is a worse problem than 1QpHab because 4QpP’s C14 dating is further away in time from its presumed creation date according to his thesis – unlike 1QpHab its two sigma range is entirely within the wrong century - and also because his thesis is more ambitious and posits a time for the Scrolls creation unlike ours, that merely claims that the dates of their creation is unclear.

To circumvent 4QpP’s C14 results Doudna makes the following arguments:

“In light of the "outlier" status of the radiocarbon date of 4QpPsa—4QpPsa gave the latest, unrechecked date of all 19 dates (the actual latest turning out to be contaminated, when rechecked)—it is simply wrong to claim that the radiocarbon date for 4QpPsa proves true dates of Qumran cave texts as late as the 1st century CE. (Remember the Bar Kochba text example above.)�

Readers of my article will note, perhaps with some amusement, that these are the exact points I make in regard to 1QpHab. That it is an outlier, that it might be contaminated, that C14 analysis is too imprecise to conclude a discussion over time spans as narrow as those found in the debate over when the Scrolls were written, and that the control sample – texts whose creation date was known - were off by over a hundred years. It will be interesting to see if my accuser – a self described “crap detector� who has made nearly 5000 posts - upon finding out that Doudna and I not only used the same C14 results but also made the same arguments regarding the place C14 dating has in the Scroll debate, will have a light go off in his head or at least hear a lever flush.

Joe

spin

John Deere,

Thank you for your admission regarding my representation of the Doudna data regarding pHab. When I had the article in my hands and gave you the data, you simply didn't believe me. You accused me of giving 1-sigma data, which is interesting because I had the paper in my hand at the time (I don't now) and could see the data which I quoted to you. That's only really taken several months for you to get here.

Afte asking you more times than I can remember for a serious justification of the figures you provided for pHab, I asked were you fabricating the data. It's a reasonable question, given your less than forward approach to your data. I didn't conclude that you had at all. I just wanted the data. You still haven't provided it. And if you feel so offended, I'd look over your previous posts for tone and think again.

What started out as a simple question, ie where exactly do your pHab datings derive, has been a serious of efforts to get past your total refusal to answer the question. You put up a smoke screen list of articles which bore no relevance except possibly the last, and when asked how exactly that article helped you to date pHab, you were once again totally unforthcoming. You started in a totally unco-operative manner and stalwartly refused to give your data source from beginning to end. In fact you still haven't given it. I ducked out of the concersation because of your unco-operative approach and only came back in when you made your ludicrous accusation: "the ludicrous conclusion".

You had not provided your data set. You just gave a date range for pHab which did not match the published range found in Doudna and which I have asked you to justify so often I'm suffering from repetitive stress. You have now indicated that you used the same data set as Doudna. Did you do the calculations?

There is little trouble with pPs37. It was worked on by John Allegro who has written that he used castor oil on his scrolls in order to clean them and make the texts clearer. Modern carbon makes texts seem younger.

It is extremely difficult to call pHab an outlier when it is with many texts in the 1st c. BCE.

Contamination almost exclusively makes texts appear younger than what they really are.

This is the most favourable of your arguments as I pointed out in the first round of this drawn-out affair. However, C14 analysis per se is not the culprit. C14 with its companion field dendrochronology have proven that extremely accurate datings are quite possible. The problem comes that various contaminants have been used on the scrolls all of which will skew the dates towards more recent times. These contaminants are not homogeneous intheir usage so they will help to cause varied datings from different parts of scrolls. Obviously the older datings they samples provide have more likelihood due to modern contaminants.

Scatology, John Deere! How endearing. If stuff sticks to you, I hope you'll wear it well.

As I pointed out from the beginning Doudna used data based on a curve announced in 1997. You were simply unaware that the data was available in 1997 -- even though I cited the fact that the data was available in 1997, giving you Doudna's statement on the matter -- and that you used was the same data. You simply ignored the indications I gave you. And now all you can do is make a scatological allusion. It's understandable.

The only reason why I came back into this discussion was due to the misrepresentations of "the ludicrous conclusion" post. Now that that has been resolved. I leave once again hoping that you don't take my name in vain again. Thank you.


spin