Table 1.

SARS-CoV-2 seroprevalence estimates from bivariate probit models with different sets of individual characteristics for non-response correction.

Regressors included in bivariate probit model	CMIA				ELISA
	Number of participants		Seroprevalence (95% CI)		Number of participants		Seroprevalence (95% CI)
	Interviewed	Tested	Naïve	Single imputation	Interviewed	Tested	Naïve	Single imputation
Demographic characteristics	6400	1038	9.0% (7.2–10.8)	8.7% (7.0–10.5)	6397	1035	10.5% (8.6–12.4)	10.1% (8.3–12.0)
Demographic and socioeconomic characteristics	6063	999	9.2% (7.4–11.1)	9.0% (7.0–11.0)	6061	997	10.8% (8.8–12.7)	10.7% (8.6–12.9)
Characteristics associated with seropositivity	6267	1026	9.0% (7.2–10.8)	7.1% (5.6–8.7)	6264	1023	10.5% (8.6–12.4)	8.6% (6.9–10.3)
Demographics, socioeconomic status and characteristics associated with seropositivity	5953	990	9.2% (7.4–11.1)	7.4% (5.7–9.2)	5951	988	10.8% (8.8–12.7)	9.1% (7.2–10.9)

“Demographic characteristics” means the following variables: individual age group (18–34, 35–49, 50–64, 65+ years old) and sex. “Socioeconomic characteristics” means the following variables: higher education status and higher self-reported income level. ”Characteristics associated with seropositivity” means the following variables: history of illness in the last 3 months, history of COVID-19 testing, whether respondent lives alone, change in hand washing habits during pandemic, week of the phone interview, and city district. All models include a variable indicating random offer of taxi transportation to and from the clinic test site for interviewed participants. All estimates are corrected for tests characteristics (see Statistical appendix for details).