Table 3. Results of the linear mixed-effects models of maintained attention towards the bought product during unplanned purchases.

	fixation duration (s)			visit duration (s)
fixed effects	estimate	SE	Χ² (p)	estimate	SE	Χ² (p)
intercept	3.09	1.37	4.58 (0.032)	3.67	0.63	5.32 (0.021)
BMI (exact)	-0.02	0.05	0.19 (0.658)	-0.03	0.06	0.25 (0.620)
energy density (exact)t	-0.001	0.001	5.39 (0.020)	-0.001	0.001	4.19 (0.041)
companion (partner)	-1.21	0.54	4.58 (0.032)	-1.21	0.62	3.65 (0.056)
gender	-0.76	0.52	2.01 (0.156)	-0.69	0.62	1.20 (0.272)
random effect	variance	SD		variance	SD
participant	0.47	0.69	-	0.40	0.63	-
residual	0.71	0.85	-	1.60	1.27	-
	fixation counts			visit counts
fixed effects	estimate	SE	Χ² (p)	estimate	SE	Χ² (p)
intercept	10.19	4.74	4.16 (0.041)	3.66	1.16	8.40 (0.004)
BMI (exact)	-0.11	0.19	0.33 (0.564)	-0.02	0.05	0.26 (0.611)
energy density (exact)t	-0.004	0.002	4.29 (0.038)	-0.001	0.001	2.49 (0.115)
companion (partner)	-2.94	1.90	2.33 (0.127)	-0.71	0.47	2.23 (0.135)
gender	-1.10	1.86	0.35 (0.556)	-0.76	0.47	2.43 (0.119)
random effect	variance	SD		variance	SD
participant	4.72	2.17	-	0.24	0.49	-
residual	11.78	3.43	-	0.87	0.93	-

Note: For each eye tracking variable a separate mixed effect model has been conducted. Gender: 0 = male, 1 = female. Presence of a companion: 0 = none, 1 = partner. P-values were calculated by likelihood ratio test of the full model against the model without the effect that has been inspected. Each p-value was calculated as likelihood ratio test of the full model against the model without the according effect. P-values are not given for covariance parameters. ^tSince estimates and standard errors for energy density are very small they are rounded to three decimal places instead of two.