Reader Comments

Post a new comment on this article

Correction

Posted by lucaturin on 11 Feb 2013 at 22:34 GMT

We are very grateful to Dr Joshua L Cherry for pointing out that we used the wrong statistic in assessing the significance of our data.

The correct one, binom.test in R [ http://www.R-project.org ] gives the following values:

Table 1
Individual p values: TG: 0.4875; DR: 0.5500; TS: 0.4194; AD: 0.4812; SG:0.5438
Aggregate probability : 0.53

Table 2
Individual p values: MG: 0.0215; LT: 0.0005; KF: 0.2891; KM: 1.52 10-5; NH: 0.0005; CC: 0.146; AM: 0.0034; CS: 0.0063; JB: 0.0005; VC: 0.1460; AD: 0.0386
Aggregate probability : < 2.2 10-16. The calculated p bottoms out at this value, presumably due to underflow, at 112/132 correct. The actual probability at 119/132 correct is therefore lower still.

Conclusions: The results with acetophenone are indistinguishable from chance, those with cyclopentadecanone are highly statistically significant. Those were the main findings of the paper, and they remain unchanged.

No competing interests declared.