TABLE 3.

Choosing the optimal amount of principal components for the squat kinetic datasets.

PC	Eigenvalue	Percentage	Cumulative	Rank of	Equality
		of variance	variance	roots	of roots
1	204.85	33.80	33.80	0.001	0.001
2	85.11	14.04	47.85	0.001	0.001
3	71.98	11.88	59.73	0.001	0.001
4	58.21	9.61	69.33	0.001	0.001
5	46.81	7.73	77.06	0.001	0.001
6	31.26	5.16	82.22	0.001	0.001
7	18.13	2.99	85.21	0.001*	0.001
8	15.18	2.50	87.71	1	0.001
9	13.82	2.28	89.99	1	0.001
10	9.55	1.58	91.57	1	0.001
11	8.38	1.38	92.95	1	0.001
12	6.37	1.05	94.00	1	0.001
13	5.57	0.92	94.92	1	0.001
14	4.65	0.77	95.69	1	0.005*
15	3.97	0.66	96.35	1	0.078

Type I error probability is set to 0.05. Rank of roots measure suggests that seven principal components (PCs) are statistically significant in meaningfully describing the dataset, corresponding to 85% of data variance, whereas the equality of roots suggests that 14 PCs are to be included (representing 95.7% of data variance). *p < 0.05.