Table 1.

Variables included in the multinomial logistic regression models.

Variable	Definition	Measurement	Link to conceptual framework
Yi	Change in consumption frequency	(1) “increase”	Food-related
		(0) “no change”	behavior
		(−1)“decrease”
x1i	Changes in food shopping frequency	Continuous variable: sum scale of changes in shopping frequency in four food categories (fresh fruit & vegetables, fresh meat & fish, other fresh food, non-fresh food)	Food-related behavior
x2i	COVID-19 risk perception	Continuous variable: Sum scale of three items measured on a 5-point interval scale	Personal factor
x3i	Closure of physical workplace	(1) “not affected”	Micro-context
		(2) “affected”: respondent's physical workplace was closed during the first lockdown
x4i	Closure of work canteens	(1) “not affected”	Micro-context
		(2) “affected”: respondent had eaten there at least once a week before the pandemic but not during the first lockdown
x5i	Closure of cafés and restaurants	(1) “not affected”	Macro-context
		(2) “affected”: respondent had eaten there at least once a week before the pandemic but not during the first lockdown
x6i	Income loss due to pandemic	(1) “not affected”	Personal resources
		(2) “affected”: Respondent had lost income due to the pandemic
x7i	Household composition	(1) “household with children”	Micro-context
		(2) “single-person household”
		(3) “household with 2+ adults, no children living in the household”
x8i	Gender	(1) “female”	Personal factor
		(2) “male”
x9i	Education	(1) “lower secondary or equivalent”	Personal factor
		(2) “upper secondary or equivalent”
		(3) “university or higher degree”
x10i	Age group	(1) “19–35 years of age”	Personal factor
		(2) “36–49 years of age”
		(3) “50–65 years of age”
		(4) “66+”
x11i	Eating frequency before pandemic	Continuous variable: 6-point interval-scale	Food-related behavior