But Boisjoly rejected Tufte's analysis
http://www.onlineethics.org/CMS/profpractice/exempindex/RB-i...
http://www.onlineethics.org/CMS/profpractice/exempindex/RB-i...
Tufte's pictures of the (superfluous) rocket graphics showing the temperature vs. his redrawing, placing temperature on the x-axis are very convincing.
1) Tufte had temperature datapoints for the previous launches that the Boisjoly's team didn't know[1]. I assume "know" to mean the complete historical data wasn't all consolidated conveniently at his fingertips before the disaster. Presumably, Boisjoly's team could have gotten it if it had occurred to them to gather it. Essentially, Tufte had the benefit of "hindsight" to make his compelling diagram showing cause & effect.
2) GIGO garbage in is garbage out. The Tufte temperature datapoint assumes that the outside ambient temperature is equal to the O-ring temperature[2] so substituting one for the other is wrong. (E.g. It's wrong to put them both on the same X-axis.)
Quotes excerpted from http://www.onlineethics.org/CMS/profpractice/exempindex/RB-i...:
[1]"He thus supposes that they knew the temperatures at launch of all the shuttles and, assuming they acted voluntarily, infers they were incompetent."
[2]", in addition, mixes O-ring temperatures and ambient air temperature as though the two were the same."
For (1), you're correct that the engineers did not have the historical data. More than that, though, it did occur to them to request the data, but they were stifled by Morton Thiokol Management (and NASA).
To start with, there were a variety of previous problems with the O-rings caused by variables that appear unrelated to temperature. These problems were resolved, but prevented anyone from seeing a pattern. It was only on the basis of the single data point in SRM 15 that Boisjoly requested temperature data in advance of the launch.
Obtaining such data was far from simple, because, as you mention, the temperature of the O-ring isn't the same as the ambient air temperature. Thus, obtaining the data was relatively involved and required knowing many variables: time on the pad, the gradient of ambient temperature, the temperature at which testing was conducted, and so forth. For this reason, the engineers didn't compile the data themselves (unclear what process they'd need to get the data).
The engineers thus requested the data in advance, but had not received it. They had precise data on only two data points (at 53 degrees and 75 degrees), so the rest of the data in the chart was compiled after the fact.
"The data necessary for a calculation of O-ring temperatures was thus not collected all along during the shuttle history. And when Boisjoly asked for that data in September, along with much other data, any one of which might have been the crucial missing piece to explain the anomalous cause, it was not supplied. In fact, the engineers received none of the data they requested."
* Both axes of the chart are inaccurate. The temperature on the X axis intermingles ambient and O-ring temperature, but they're not the same (imagine ocean temperature vs air temperature). The vertical axis, O-ring damage, is semi-relevant, but the important question is whether the O-rings held a seal (degree of blow-by). O-ring damage probably correlates, but seems like a proxy that may be misleading.
* The data available in the chart was not available to the engineers at the time. There was only one previous failure (SRM 15) that led anyone to believe temperature itself was an issue. They had two valid temperature data points (SRM 15 and SRM 22), and correctly pointed out that another launch at 29 degrees was completely outside their tested range and would be inherently risky. It's not clear to me that a scatterplot with two datapoints would be that effective.
Furthermore Boisjoly criticizes that Tufte presumed to judge the engineers, for the following reasons:
* The engineers previously recommended no launch should occur, months in advance (due to previous O-ring issues), but NASA overruled them.
* The engineers recommended, and their managers accepted, that the low launch temperature was outside their test database. NASA came back and requested proof that the temperature was dangerous, but the engineers could not comply, because the parameters were extreme enough that they had not tested them. They couldn't prove a negative. The managers at Morton Thiokol and NASA then jointly overruled the engineers.
* The idea of putting together a chart was not something the engineers considered, because there were a variety of previous unrelated problems in O-rings (each resolved afterwards) that muddied that data, and muddies the data that Tufte presents.
* Tufte himself failed to research or note the pivotal information, and thus misrepresented the situation the engineers were in, and the data available, while imputing that the engineers should be held morally responsible for not presenting data they didn't have (and then going on to present such data himself, compiled after the fact and at his own leisure, incorrectly).