Hand Tools Archive

Re: Survey information and what you can surmise

david weaver
There is more than one chart from Rob's data experiments, but I recall one of the early charts showing a rounded figure of 60 for V11. The published range for V11 having looked yesterday is 61-63, and anyone who deals in likelihoods would say the following things are possible:

* the chisel that rob tested was 61 hardness or more and rob's reading was approximately 60 or 59.x, whatever it was (this seems the most likely)

* the chisel that rob tested was approximately 60 hardness despite most chisels being harder, and rob's measurement was accurate (this seems less likely, thus we'd love to have more than one person do this type of experiment to see what the data says - in the event that a spec is 61-63, if something ever does slip out from someone like a LV or LN, it wouldn't be the norm)

* that LV's found the bottom end of the spec to make a more workable chisel and some of the tools are regularly outside of it without changing the published spec. (what we know of LV makes this near zero chance)

Those are just my opinions.

When someone takes on a project like this on their own accord and publishes data, there isn't really an expectation of resolution vs. accuracy being lab standards (that is, you're suggesting that the display of two decimal places implies that level of accuracy - that's fine in peer reviewed work, but even in that case, standards are often blown). I think a reasonable individual would look at a data set like this and compare it to experience and ask how big the error could be. i doubt most of the public is familiar with the issue of resolution vs. precision in presenting data, even if they understand generally what c-scale hardness numbers represent.

That leaves us with two options - discuss data provided on forums not at all unless it's peer reviewed, or discuss them as they are and see if we can glean anything from them. I'd prefer the latter as long as absolute assertions aren't made (as in, you cannot just look at this data set and say "well, LV's chisels are softer than their spec" based on one test, and you can't say "look, blue spruce makes 65 hardness butt chisels").

Even better than either of these (no information or errant information) would be additional efforts that are more precise and controlled (and with a larger sample of examples). Once in a while, someone comes along and tests things for no obvious reasons, but it's rare.

© 1998 - 2017 by Ellis Walentine. All rights reserved.
No parts of this web site may be reproduced in any form or by
any means without the written permission of the publisher.