. 2021 May 26;16(5):e0251571. doi: 10.1371/journal.pone.0251571

Table 6. The estimated probabilities of behaviour related to either the dog (D and DA visits) or the activity (DA and A visits).

Behaviour variable	Period			Statistics
	1	2	3
Physical contact to dog (D, DA)	0.37^c [0.23; 0.59]	0.49^b [0.33; 0.58]	0.52^a [0.35; 0.74]	1 vs 2: n = 95, S = 555 P < 0.05
				1 vs 3: n = 94 S = 1008 P < 0.001
				2 vs 3: n = 94, S = 670 P < 0.01
Talk directed at the dog (D, DA)	0.92^a [0.84; 0.95]	0.90^ab [0.73; 0.95]	0.92^b [0.68; 0.96]	1 vs 3: n = 94, S = -546 P < 0.05
Talk about the dog (D, DA)	0.25^b [0.13; 0.50]	0.31^a [0.17; 0.55]	0.34^ab [0.19; 0.48]	1 vs‥ 2: n = 90, S = 857 P < 0.001
Physical contact to activity (DA, A)	0.60^b [0.38; 0.75]	0.61^a [0.49; 0.72]	0.70^a [0.36; 0.90]	1 vs 2: n = 93, S = 775 P < 0.01
Physical contact to activity (DA, A)	0.60^b [0.38; 0.75]	0.61^a [0.49; 0.72]	0.70^a [0.36; 0.90]	1 vs 3: n = 91, S = 543 P < 0.05
Talking about activity (DA, A)	0.57^b [0.25; 0.81]	0.84^a [0.45; 0.94]	0.72^a [0.43; 0.92]	1 vs 2: n = 93 S = 1801 P < 0.0001
Talking about activity (DA, A)	0.57^b [0.25; 0.81]	0.84^a [0.45; 0.94]	0.72^a [0.43; 0.92]	1 vs 3: n = 88, S = 793 P < 0.001

The estimated probabilities are shown as median and interquartile range (IQR) for each period. The estimates derive from the final generalised linear mixed models, and are corrected for significant factors in the respective models. Letters within variable (a, b) indicate differences across period. Visits with a dog, no activity (D); Visits with a dog, with activity (DA); Visits without dog, with an activity (A).