Table 3.

Results of statistical analyses (t-test, Mann–Whitney rank sum test, and Dunn’s method performed after Kruskal–Wallis one way ANOVA on ranks) of traction force values obtained on different samples.^a

comparison	d.f.	test statistics	P	significantly different
glass: 1 vs 2
C36	38	t = 0.048	0.962	no
C40	—	T = 428	0.636	no
C44	38	t = 0.580	0.565	no
C50	—	T = 4040	0.862	no
glass vs wax
C36	—	Q = 7.203	<0.05	yes
C40	—	Q = 7.743	<0.05	yes
C44	—	Q = 6.938	<0.05	yes
C50	—	Q = 5.734	<0.05	yes
between wax
C36 vs C40	38	t = 1.105	0.312	no
C36 vs C44	—	T = 396	0.715	no
C36 vs C50	—	T = 319	0.015	yes
C40 vs C44	—	T = 362	0.198	no
C40 vs C50	—	T = 299	0.003	yes
C44 vs C50	—	T = 340	0.062	no

^ad.f., degrees of freedom; P, probability value; Q, Dunn’s method statistics; t, t-test statistics; T, Mann–Whitney rank sum test statistics; yes, significantly different; no, no significant difference.