Continuing the discussion from Some kind of waterfall that quickly fossilized teddy bears:
Yes Craig, I will help you. You need to understand the basic principles of how measurement works before you try to discuss the reliability or otherwise of radiometric dating. Allow me to explain a couple of important facts to you.
1. Unreliability is always quantified.
You said:
That is a straw man. Nobody is claiming that the assumptions behind radiometric dating are always correct. No form of measurement of any kind works like that. Every form of measurement has potential sources of error, even the tape measure that I used today to measure my desk. But – and it is a big but – there is a massive difference between “occasionally out by a few percent” and “consistently out by a factor of a million.”
In every area of science, researchers are required to identify, catalogue, quantify, and account for all possible sources of error. This means that, among other things, they have to determine the maximum extent to which their assumptions could have been violated. Geochronologists are fully aware that there could be contamination or leakage. But they take steps to figure out exactly how much contamination or leakage there could have been, and what effect it could have had on the overall result. So when they tell us that the Earth is 4.54±0.05 billion years old, we can be confident that they have taken these uncertainties into account.
Scientists generally divide sources of error into two categories: random errors and systematic errors. Random errors represent the “spread” of a number of measurements of the same thing around an average value, and that is what is represented by the error bars. Systematic errors represent an offset from zero that applies to all measurements across the board in this particular case.
Your Novarupta example, assuming that it was collected and processed correctly, is an example of a systematic error. It indicates that that particular method might overestimate ages by up to 5.5 million years. This means that Precambrian rocks dated by the same method to 550 million years could be as young as 544.5 million years or possibly at a stretch 540 million years. It does not justify claims that they could be as young as six thousand. You CANNOT claim that a measurement technique is significantly more unreliable than what is indicated by the random and systematic errors.
2. Unreliability is specific.
Another thing you need to understand is that radiometric dating is not just one single technique, but a whole range of related techniques. There are about forty different isotopes used in radiometric dating, from carbon-14 through to uranium/thorium/lead via potassium-argon and rubidium-strontium. Each of these will have different considerations, different preconditions, different chemical and physical properties, and different subtleties. There are also dozens of different minerals that may or may not be able to be dated radiometrically. Some of them will be more reliable than others. Some of them may overestimate ages if their assumptions are not met, but others may underestimate them.
This means that you have to consider each different case individually. You cannot make any blanket claims that radiometric dating as a whole is “unreliable” just because some methods give flaky results on some minerals. K-Ar dating may give flaky results due to the fact that argon is a gas and more likely to leak out or be trapped in lava as it solidifies, but that tells us nothing whatsoever about the reliability of U-Pb dating of zircon crystals.
This is a point that I keep having to make over and over again in response to young Earth claims, Craig. It is one thing to show that some methods give errors of up to a few million years on some minerals under some circumstances. It is a completely different matter to show that all methods are consistently overestimating ages by up to a few billion years – a thousand times as much – in every circumstance. Such a leap is simply not justified. There is nothing “evolutionist” or “biased” about me saying so – that is simply how measurement works in every area of science.