*3.7. Criteria*<sup>18</sup>

An unaddressed issue concerns the setting of criteria with respect to closeness and the probability of being close. The temptation is to set cuto ffs; for example, that closeness must be at the 0.02 level or better, at 95% probability or better, to justify publication. However, this temptation must be resisted, lest the APP degenerate into dichotomous thinking that is currently problematic in the sciences. Instead of cuto ffs, it would be better to have graduated verbal descriptions for what constitutes di fferent levels of closeness and probabilities of being close. However, even graduated verbal descriptions may be problematic because researchers in di fferent fields, or areas within fields, might justifiably di ffer with respect to what constitutes suitable verbal descriptions. For example, closeness at the 0.40 level might be "poor" in some fields or areas, and "acceptable" in others. It would be a mistake to impose such criteria from outside.<sup>19</sup>

One way to address the issue would be to have conferences, workshops, or symposia where people in similar fields and areas meet; become familiar with the APP, with a solid understanding about closeness and the probability of being close; and engage in serious discussions about criteria for

<sup>18</sup> I thank an anonymous reviewer for suggesting this issue.

<sup>19</sup> Trafimow (2018a) used some graduated verbal descriptions but also emphasized that these should not be taken very seriously.

graduated verbal descriptions. An alternative possibility would be to use the social media. Yet another alternative would be for editors of substantive journals to promulgate special issues devoted to setting criteria for graduated verbal descriptions such that substantive experts can provide perspectives.

Given academia's publish or perish culture, it cannot be overemphasized what a mistake it would be to turn APP thinking dichotomous, with publication thresholds for closeness and probability of being close. The ability of researchers to meet criteria for various verbal descriptions should only be one consideration for publication. Many factors should influence publication decisions, including the worth of the theory, the execution of the study, the feasibility of obtaining participants, the writing clarity, and others. Hopefully, having graduated verbal descriptions, instead of dichotomous cutoffs, that differ across fields and areas; will facilitate journal editors and reviewers to engage in more nuanced thinking that weighs many relevant factors.
