Well, a little color. As covered in It was an itsy-bitsy teenie weenie yellow polka dot error, Kirk Shanahan digitised a chart from page 87 of Storms, The Science of Low Energy Nuclear Reaction, even though the data was on the next page in Table 7. Ah, well, you do what you need to do.
So, today, I loaded the data in to a spreadsheet, and here it is, ODS, and if you need another format, ask. The first plot shows all the data, and looks like the Storms plot, but with a little extra and without the 23.8 MeV/He line; that is equivalent to about 2.6 x 10^11 He atoms/watt-sec.
What is this showing? Without getting into all the details for the Miles experiment, what is plotted is measured anomalous power (excess heat) for a collection interval vs the calculated helium atoms per watt-second. With the Miles data, the measurement background of 0.51 x 10^14 atoms per 500 ml (the collection volume) is subtracted from the measured helium. We can see an obvious outlier, and that was at the lowest measured excess power (20 mW), where we would expect error to loom larger. While there is obvious and substantial scatter in the Miles data, it is useful to keep in mind that any correlation at all was considered a major discovery in 1991, when this was announced, and that the reported ratio was within an order of magnitude of the theoretical d-d fusion -> helium yield was considered amazing (by Huizenga, of all people). (Miles’ earliest announcement had order-of-magnitude helium results, later, his helium measurements were improved).
To get a better idea, I eliminated the obvious outlier (as Storms does in calculating the average on his next page). That produces this chart:
What we can see is how the Bush and Lagowski data fits into this as having improved precision.
One basic problem in interpreting these results is that it is possible that the retention ratio varies with cathode surface conditions, and that could vary even within a single experiment. (i.e., we should understand that though ten results from Miles are presented, there were not ten separate experiments. These were distinct collection intervals, power being measured during a collection interval, with collection ceasing when 500 ml has been collected.
(I am writing this from memory, and corrections will be appreciated). We know that surface conditions change as a cathode is conditioned by loading and deloading, and this is poorly controlled. Electrolysis is messy.)
What is measured here is helium in the outgas, not total helium production, so even perfect measurement might still show substantial variation. Hence the goal becomes to measure all the helium, and there are only two experiments where that may have been done, SRI M4 and Apicella et al Laser-3; these are the only known experiments where anodic erosion was used (by reversing electrolytic polarity), and the amount of helium found in both of them moved up to within experimental error of 100% of the theoretical value (23.8 MeV/4He or 2.6 x 10^11 4He/watt-sec.)
As can be seen in the charts, scatter does lessen as the measured quantities increase.