Section 1.1, page 10: “He calculated that there would be an arctic temperature increase of approximately 8º C (46.4º F) from atmospheric carbon levels two to three times their known value at the time.”
While a temperature of 8º C converts to a temperature of 46.4º F, an 8º C change in temperature (as in this sentence) does not. An arctic temperature increase of 8º C would convert to an increase of 14.4º F.
Thank you to Dana Tulodziecki for pointing out this conversion error!
Section 10.3, page 259: "Increasing sample size to increase the power of a statistical test can be a good thing, but it also has a downside: this increases the chance of making a type I or false positive error."
This is incorrect as written. Thank you to Julia Rohrer for pointing this out!
Consideration of the paragraph can help clarify what we had in mind:
“A related issue is that statistical tests vary in their power to detect an effect. The power of a statistical test is the probability that the test will enable the rejection of a null hypothesis. More powerful tests increase the chance of rejecting the null hypothesis, thus decreasing the chance of a type II or false negative error, where we fail to reject the null hypothesis when it is actually false. Power increases with sample size. In the tea-tasting experiment, we weren’t able to reject the null hypothesis after one cup was guessed correctly, but we were able to after eight cups were guessed correctly. Increasing sample size to increase the power of a statistical test can be a good thing, but it also has a downside: this increases the chance of making a type I or false positive error. Studying a very large sample makes it relatively easy to uncover statistically significant findings, but this also makes it relatively easy to erroneously reject the null hypothesis—that is, to uncover findings that turn out to be false.”
It is a mistake to say that a large sample size impacts the rate of type-I error (which is fixed by a choice of alpha level). But, in the tea-tasting experiment described here, the sample size is not chosen in advance. When an experiment iteratively tests for significance as data accumulate, the risk of false positives does become greater. So, what should be said is that a large sample size without a pre-registered analysis plan can increase the risk of making a false positive error.