Skip to main content

Table 5 Analysis 2 (German only): model comparison

From: Capturing multiple interaction effects in L1 and L2 object-naming reaction times in healthy bilinguals: a mixed-effects multiple regression analysis

ModelsInformation criteria (log likelihood)Deviance (− 2* log likelihood)Number of parametersChi-square obtaineddfp-value
Fixed EffectI only10,363.62− 20,704.642   
Plus subjI10,509.75− 21,019.503292.271p < 0.0001
Plus subjI. ItemI10,589.46− 21,178.924159.411p < 2.2e−16
Plus runNum10,671.27− 21,342.545163.631p < 2.2e−16
Plus trialNum10,686.06− 21,372.12629.561p = 5.416e−08
Plus log10WF10,697.36− 21,394.72722.611p = 1.986e−05
Plus firstPhonemeDiff10,704.05− 21,408.10813.391p = 0.000253
Plus wordChoice10,707.79− 21,415.5897.461p = 0.006300
Plus GermanRun1PercentCorrect10,710.99− 21,421.98106.401p = 0.011383
Plus gender10,713.03− 21,426.06114.091p = 0.043243
Plus gender*firstPhonemeDiff10,715.07− 21,430.14124.081p = 0.043447
Plus GermanRun1PercentCorrect*log10WF10,717.46− 21,434.92134.781p = 0.028846
  1. This table shows a comparison of different statistical models used for the within-German reaction time (RT) comparisons