Skip to main content

Table 4 Explained variance of clusters of factors using hierarchical mixed modelling

From: Explaining variance in perceived research misbehavior: results from a survey among academic researchers in Amsterdam

  Perceived frequency of misbehavior Perceived impact of misbehavior
Clusters added1 Index of expl. var0 Cum.explained variance1 Diff. model fit2 (df) Signif. model fit3 Expl. var0 Cum. explained variance1 Diff. model fit2 (df) Signif. model fit3
Individual factorsa 6.74% 6.74% 74.1 (6) <.001 1.18% 1.18% 18.1 (6) .001 < p < .01
Climate factorsb 22.22% 31.64% 358.2 (7) <.001 14.10% 15.66% 205.7 (7) <.001
Publication factorsc 15.85% 34.21%* 32.5 (3) <.001 12.28% 18.42%* 37.6 (3) <.001
  1. 0 = this is the explained variance when only one group of factors is analyzed, i.e. just the climate factors explain 22.22% of variance perceived frequency of research misbehaviors
  2. 1 = the explained variance here is the cumulatively explained variance. Since the models are hierarchical, factors are added consecutively, i.e. the explained variance is 31.64% when both individual as well as climate factors are added to the model
  3. 2 = difference in model fit, model fit here is the difference between the − 2 Log likelihood of the previous model, i.e. 74 is the difference between the intercept-only model and the model with individual factors added, etc.
  4. 3 = significance of model fit is contrasted with the previous model; the row above or a model with no parameters (vs. individual factors)
  5. a = gender, academic rank and disciplinary field, b = SOURCE subscales, c = PPQr subscales
  6. * Note that the explained total is less than the sum of its parts because there is some overlap in the variance that publication and climate factors explain