. Author manuscript; available in PMC: 2021 Aug 1.

Published in final edited form as: Emotion. 2019 Jun 13;20(5):773–792. doi: 10.1037/emo0000609

Table 2.

Results of null, polynomial, and thin plate smoothing spline models for each dependent variable.

Dependent variable	Model	β	p	Adj R²	AIC

Emotion comprehension scores	Null				1608.10
	Linear	.48	< .001^***	.23	1558.51
	Quadratic	−.50	< .001^***	.47	1484.31
	Cubic	.32	< .001^***	.58	1442.01
	Spline			.62	1426.36
Coder abstractness scores	Null				833.48
	Linear	.76	< .001^***	.58	663.07
	Quadratic	−.37	< .001^***	.72	586.51
	Cubic	−.01	.745	.72	588.40
	Spline			.73	580.05
Linguistic abstractness scores	Null				−142.50
	Linear	.46	< .001^***	.21	−185.25
	Quadratic	−.001	.991	.20	−183.25
	Cubic	−.05	.467	.20	−181.79
	Spline			.21	−185.25⁺
General definition use	Null				1859.18
	Linear	.70	< .001^***	.49	1728.89
	Quadratic	−.46	< .001^***	.70	1627.36
	Cubic	.07	.057	.70	1625.66
	Spline			.70	1623.66
Synonym use	Null				1738.33
	Linear	.32	< .001^***	.10	1719.66
	Quadratic	−.05	.458	.09	1721.10
	Cubic	−.01	.848	.09	1723.06
	Spline			.10	1719.66⁺
Example situation use	Null				1866.19
	Linear	−.67	< .001^***	.45	1750.57
	Quadratic	.32	< .001^***	.55	1712.61
	Cubic	.02	.687	.55	1714.44
	Spline			.56	1710.20
Physiological marker use	Null				1392.78
	Linear	−.23	.001^**	.05	1384.27
	Quadratic	−.01	.882	.04	1386.25
	Cubic	.13	.062	.05	1384.68
	Spline			.10	1378.66

Note: Bold text indicates best fitting model for each dependent variable, as determined by having the lowest AIC.

⁺

In these two cases, spline model algorithms revealed that the best fits were identical to linear models; hence we interpreted the linear models. β = standardized beta, Adj R² = adjusted R², AIC = Akaike Information Criterion,

^***

p < .001,

^**

p < .01