Table 4 Tested models for recollection.

From: The impact of musical pleasure and musical hedonia on verbal episodic memory

Fixed effects

df

Measures of fit

Likelihood ratio tests

AICci

∆iAICc

wiAICc

LLi

\({x}^{2}\)

df

p

Cond*BMRQ*Back

13

12,570

9.01

0.01

− 6272

   

Cond*BMRQ + Back

8

12,563

1.33

0.32

− 6273

2.34

5

0.801

Cond*BMRQ

7

12,561

0.00

0.63

− 6274

0.67

1

0.412

Cond + BMRQ

5

12,567

6.23

0.03

− 6279

10.24

2

0.006

  1. All models included the Participant variable as random intercept. In the formulas, ‘Cond’ = Condition, ‘BMRQ’ = BMRQ score, ‘Back’ = Background, and ‘*’ = interaction. ‘df’ = degrees of freedom. AICci = corrected Akaike Information Criterion. ∆i(AICc) = difference between AICc for model i and best model’s AICc. wi(AICc) = Akaike weight measuring the level of support in favour of model i being the most parsimonious among the candidate model set. LLi = natural logarithm of the maximum likelihood for model i. Likelihood ratio tests compare the goodness of fit of that particular model to the previous one. Backward elimination was performed until no further variables or interactions could be removed without a statistically insignificant loss of fit.
Back to article page