Table 1.

Pearson correlations between the semantic questions trained to the variable age.

Label	r	p	r2	RMSE	Min	Max
All texts	0.310**	<0.0001	0.096	12.03	35.65	62.25
Sleep	0.191**	<0.0001	0.036	14.08	35.46	103.19
Self	0.170**	<0.0001	0.029	11.73	35.50	46.58
Affect behaviour	0.151**	<0.0001	0.023	15.48	35.79	83.53
General reason	0.137**	<0.0001	0.019	11.93	35.46	49.97
Energy	0.122**	<0.0001	0.015	11.93	35.48	49.18
Harmony	0.116**	0.0003	0.013	11.98	35.51	47.86
Depression	0.098**	0.0019	0.010	12.44	35.42	54.63
Movement	0.087*	0.0054	0.008	12.81	35.43	70.70
Worry	0.083*	0.0072	0.007	15.72	34.77	89.05
Depression, reason	0.072*	0.0161	0.005	12.36	35.46	72.90
Worry, reason	0.071*	0.0181	0.005	13.24	35.33	58.09
Satisfaction	0.069*	0.0202	0.005	12.56	35.17	70.00
Concentration	0.065*	0.0272	0.004	12.15	35.51	53.12
Appetite	0.005	0.4403	0.000	13.04	35.63	60.71
General	−0.007	0.5822	0.000	12.48	35.59	51.91
Interest	−0.011	0.6313	0.000	13.78	35.93	92.32

*p  <  0.05 uncorrected for multiple comparisions, **p < 0.05 following Bonferoni correction for multiple comparisions.

Note. The rows show the label of the open-ended questions, the Pearson correlation coefficients (r), the p values, root mean squared error (RMSE), minimum predicted age (Min), and maximum predicted age (Max). The p value states the probability of observing a correlation at least as large (in absolute terms) as the observed correlation under the assumption that the true correlation is 0.