Removal of Nominees. We conducted the same bootstrapped significance tests as in Study 1; that is, we created a 99% confidence interval around reliability estimates and intercor- relations by resampling random nominee removal in the same pro- portion that nominators were missing in each school. The means of the bootstrapped distributions of reliability coefficients were the same as the full sample (nominators systematically removed but not nominees) in all cases to two decimals. The mean of the boot- strapped correlations were almost identical to the “nominees in” dataset, with no difference exceeding .01.
Appears in 6 contracts
Samples: core.ac.uk, repository.ubn.ru.nl, repository.ubn.ru.nl
