. 2019 Mar 12;19(5):1241. doi: 10.3390/s19051241

Table 2.

Linear mixed effect model for the drinking task at 30 cm distance.

	Side	Reliability	Cycle Time	Elbow FE	Elbow R	Shoulder FE	Shoulder AbAd	Shoulder R
	Dominant	Within-subject	0.92	0.88	0.97	0.85	0.80	0.90
		Inter-rater	0.81	0.47	0.002	0.67	0.72	0.78
G-coefficients		Overall	0.93	0.68	0	0.86	0.93	0.90
	Non-Dominant	Within-subject	0.90	0.96	0.96	0.87	0.90	0.87
		Inter-rater	0.84	0.42	0.14	0.72	0.81	0.60
		Overall	0.95	0.60	0.23	0.87	0.93	0.79
	Dominant	Intercept	5.65 s ***	90.76° ***	20.09° ***	5.68° **	14.69° ***	−5.00° **
	n = 397		(0.20)	(2.77)	(5.62)	(1.76)	(1.64)	(1.68)
		σ²_rat	0.00	0.00	0.00	0.00	0.91	0.00
		σ²_rep	0.00	2.38	0.00	0.16	0.00	1.04
		σ²_subj	0.71	100.31	0.00	53.09	41.88	48.74
Model and		σ²_{rep_rat}	0.00	0.88	0.00	0.00	0.14	0.00
variance		σ²_{subj_rat}	0.10	92.30	1260.94	16.06	4.53	10.01
components		σ²_{subj_rep}	0.01	0.00	2.08	1.22	0.57	0.77
		σ²_residual	0.06	23.05	36.25	10.46	11.02	4.44
	Non-dominant	Intercept	5.68 s ***	97.79° ***	24.75° ***	2.94°	15.84° ***	−3.15°
	n = 376		(0.19)	(2.96)	(5.38)	(2.37)	(1.74)	(1.62)
		σ²_rat	0.00	0.00	0.00	0.22	0.00	0.00
		σ²_rep	0.00	0.03	3.67	0.00	0.00	0.00
		σ²_subj	0.62	99.10	124.19	90.38	53.51	39.08
		σ²_{rep_rat}	0.00	0.40	0.00	0.82	0.09	0.21
		σ²_{subj_rat}	0.05	133.03	832.17	24.28	6.91	19.92
		σ²_{subj_rep}	0.01	1.36	9.89	4.21	0.68	1.33
		σ²_residual	0.07	7.43	23.90	11.52	5.94	7.22

*** p < 0.001, ** p < 0.01, * p < 0.05. Statistical model: Response ~ (1 | “Subject”) + (1 | “Rater”) + (1 | “Repetition”) + (1 | “Subject”:”Repetition”) + (1 | “Subject”:”Rater”) + (1 | “Repetition”:”Rater”). The dataset was stratified using the factor “arm dominance”. In the table, σ²_sub, σ²_rat, σ²_rep, σ²_{rep_rat,} σ²_{subj_rat,} σ²_{subj_rep,} and σ²_residual are the variance between subjects, among raters, among repetitions, of the interaction between repetitions and raters, subjects and raters, subjects and repetitions, and residuals respectively. Values within parentheses indicate the mean square error of each parameter estimate. The reliability estimates (defined by the G-coefficient, absolute agreement) for the different responses are subdivided into within-subject, inter-rater, and overall reliability. The number of subjects = 20 and the number of raters = 2.