Table 4 Nested model comparisons of content and dynamic models predicting brooding scores.

From: The think aloud paradigm reveals differences in the content, dynamics and conceptual scope of resting state thought in trait brooding

Nested Model Comparisons (Manually-coded content)

 

AIC

r2

Adj. r2

P

SS

∆SS

F(df)

P

Model 1 (Content + Covariates)

251

24.31%

17.73%

0.011*

323.93

   

Model 2 (add Total Word Count)

246.4

33.48%

26.08%

0.002**

284.72

39.21

7.21 (1)

0.010*

Model 3 (add Dynamics)

242

47.88%

36.44%

 < 0.001***

223.08

61.64

2.83 (4)

0.037*

Nested Model Comparisons (LIWC-coded content)

 

AIC

r2

Adj. r2

P

SS

∆SS

F(df)

P

Model 1 (Content + Covariates)

247.8

31.65%

24.03%

0.003**

292.54

   

Model 2 (add Total Word Count)

243.6

39.45%

31.20%

 < 0.001***

259.14

33.39

6.29 (1)

0.016*

Model 3 (add Dynamics)

241.5

50.38%

37.98%

 < 0.001***

212.36

46.78

2.2 (4)

0.086

  1. Nested model comparisons assessed the explained variance and overall fit of three models predicting brooding scores. Model 1 for manually-coded content (top) included the significant predictors: valence, % past-orientation, % internal-orientation, censorship, and similarity to daily life. Model 1 for LIWC-coded content included % negative words, % past-oriented words, % first-person pronouns, censorship and similarity to daily life. Model 2 was identical to Model 1 but also included the third covariate of total word count. Model 3 included the same variables as Model 2 as well as mean number of words for positive thoughts, the mean number of words for negative thoughts, and affective transition probabilities from Pos → Pos and Neg → Pos.
  2. AIC Akaike Information Criterion, r2: r squared, adj. r square: adjusted r square, df degree of freedom, SS sum of squares, ∆SS variation of sum of squares. Highlighted in bold are statistically significant models at p < .05. *P < .05, **P < .01, ***P < .001.
Back to article page