Antibodies to selected Plasmodium falciparum merozoite antigens are often reported to be associated with protection from malaria in one epidemiological cohort, but not in another. Here, we sought to understand this paradox by exploring the hypothesis that a threshold concentration of antibodies is necessary for protection. We analyzed data from two independent cohorts along the Kenyan coast, one in which antibodies to AMA1, MSP-2 and MSP-3 were associated with protection from malaria (Chonyi) and another in which this association was not observed (Junju). We used a malaria reference reagent to standardize antibody measurements across both cohorts, and applied statistical methods to derive the threshold concentration of antibodies against each antigen that best correlated with a reduced risk of malaria (the protective threshold), in the Chonyi cohort. We then tested whether antibodies in Junju reached the protective threshold concentrations observed in the Chonyi cohort. Except for children under 3 years, the age-matched proportions of children achieving protective threshold concentrations of antibodies against AMA1 and MSP-2 were significantly lower in Junju compared to Chonyi (Fishers exact test, P