Table 1 Goodness-of-fit tests of the five pair-choice aggregation models in terms of chi-square statistics.

From: Vasopressin enhances human preemptive strike in both males and females

Model

Overall

Saline

AVP

χ2

p

χ2

p

χ2

p

More aggressive wins

138.230

<0.001

94.544

<0.001

46.441

<0.001

More peaceful wins

6.651

0.036

3.441

0.179

9.219

0.010

More extreme wins (with aggression advantage)

18.198

<0.001

13.038

0.001

8.019

0.0181

More extreme wins (with peace advantage)

1.004

0.605

2.775

0.250

1.550

0.461

Random

8.676

0.013

5.292

0.071

4.098

0.129

  1. Note. To assess each model’s goodness of fit to the observed frequencies of pair choices, we conducted a chi-square test in a two-way cross-table with the model/observed × the response categories (attack, move, untouched: see Supplementary Tables S3 and S4 for the frequencies/percentages predicted by each model). Here, the 2nd column shows results of the model testing with the choice frequencies collapsed over treatments (saline vs. AVP) and sexes. The 3rd and 4th column reports a separate test for the saline and AVP condition respectively with the choice frequencies collapsed over sexes. The “more extreme wins (with peace advantage)” model was the only one that was not rejected (i.e., approximated the observed pair behavior reasonably well). In all rows, the df was 2.
Back to article page