Temperature: If one looks at annual mean temperature in isolation, it could easily yield the mistaken view that higher temperature results in a higher likelihood of malaria occurring in a country. For instance, a simple comparison of histograms showing the fraction of countries by temperature, dividing the sample into places where malaria has been eradicated (top panel) versus where it is still present (bottom panel), could lead to the inference that the higher the mean annual temperature, the greater the number of countries with malaria.
But this is incorrect because it fails to control for the influences of income, household size and other socioeconomic characteristics. The multivariate analysis shows that when these factors are controlled, higher temperatures are actually associated with a small but significant increase in the probability of malaria eradication.