Skip to main content

Table 1 Shapiro-Wilk tests for deviation from normality in the distribution of total brood weights for different colony sizes in the late spring (6 November) and early summer (4 December) samples

From: Evolution of sociality by natural selection on variances in reproductive fitness: evidence from a social bee

Date

Colony size (females/nest)

Sample size (number of colonies)

Shapiro-Wilk statistic

Significance

Re-sample comparison (Sig.)

November

1

46

0.622

<0.001

-

 

2

63

0.810

<0.001

<0.01

 

3

25

0.915

0.039

<0.01

 

4

10

0.846

0.053

0.08

December

1

31

0.728

<0.001

-

 

2

38

0.895

0.002

<0.01

 

3

9

0.918

0.375

0.04

 

4

5

0.912

0.481

0.19

  1. The re-sample comparison gives the proportion of re-sampled 1-female colony brood weights with higher Shapiro-Wilk values than for 2, 3 or 4-female colonies. These proportions approximate the likelihood that a higher Shapiro-Wilk value could have been obtained from 1-female nests given the same sample size as for the multi-female colony size (see Methods).