Table 3.

Association between testing experience and confidence in test

	Outcome: Believed Negative RDT very likely correct at second annual survey
	(1)	(2)	(3)	(4)	(5)	(6)
Coefficient on:
Any RDT	3.44** (1.40)	3.63** (1.39)
Negative RDT result			1.86 (0.65)	2.10 (0.82)
Positive RDT result					2.40* (1.07)	2.30 (1.12)
Believed Neg. RDT very likely correct at first annual survey	2.68 (1.96)	4.30 (3.92)	2.59 (1.91)	3.90 (3.50)	2.40 (1.82)	3.79(3.70)
Mean of outcome in ref. group	0.539	0.539	0.650	0.650	0.617	0.617
Number of observations	179	179	179	179	179	179

Beliefs are those of the household respondent if the individual was under the age of 18. Information on whether the individual was tested and the test result was based on sick visit surveys by the study team. Results are from logistic regression models and coefficients are expressed in terms of odds ratios. Columns1,3,5 are regressions that control only for baseline beliefs (measured at the first annual survey) while columns 2, 4, and 6 include the following controls: age and gender of the individual, education level (of the respondent if the individual was under 18), the main source of household drinking water, whether the household owns more than one acre of land and village fixed effects. *p < 0.05, **p < 0.01