Table 2 Prediction accuracies of the Self-Compassion Scale score by text.

From: Advancing psychological assessment: quantifying self-compassion through free-text responses and language model BERT

 

Question 1 (thought-general)

Question 2 (thought-self)

Question 3 (thought-others)

Question 4 (action)

Combined texts (per situation)

Situation 1 (suffer)

0.47***

0.50***

0.42***

0.38***

0.57***

[0.34, 0.58]

[0.37, 0.61]

[0.28, 0.54]

[0.23, 0.50]

[0.46, 0.67]

Situation 2 (shortcomings)

0.48***

0.57***

0.41***

0.37***

0.63***

[0.35, 0.60]

[0.45, 0.67]

[0.27, 0.54]

[0.22, 0.50]

[0.53, 0.72]

Situation 3 (failure)

0.44***

0.48***

0.44***

0.31***

0.56***

[0.30, 0.56]

[0.35, 0.59]

[0.30, 0.55]

[0.16, 0.44]

[0.44, 0.66]

Combined texts (per questions)

0.59***

0.67***

0.51***

0.52***

–

[0.48, 0.68]

[0.57, 0.75]

[0.39, 0.62]

[0.39, 0.62]

  1. Values in square brackets indicate the 95% confidence interval for each correlation. ***p < 0.001.
Back to article page