Search performance figures (recall and precision) can be treated as proportions, for the purpose of establishing confidence intervals and it can be shown that the confidence interval decreases, for a given sample size, as the true recall increases.
The first observation from Fig. 2 is the fact that the precision and recall curves do cross as expected, since the precision drops and recall increases as the absolute cutoff point increases, i.e., the considered extracted terms lists becomes less restrictive.
Jaccard et al. (2004) compared recall periods of 1, 3, 6 and 12 months, and found that, for those with multiple sex partners, recall errors in self-report of number of sex partners increased as recall period increased.
Indeed, the probability to correctly map queries from users to profile (i.e., the recall) increases as the number of possible user profiles decreases.
This may reflect an increase in variability of number of sexual partners reported as the length of recall increases.
As mFAT increases, precision declines and recall increases.
As expected, precision decreases with the distance threshold, while recall increases.
In addition, as the merg_thred grows from 0.1 to 0.9, the recall increases while the precision decreases.
Graham et al. (2003) compared recall periods of one, two, and 3 months and found evidence that, for a high frequency behavior (i.e., heterosexual vaginal sex), accuracy decreased as recall period increased.
The difference may reflect biases in obtaining the data, such as recall bias (increasing vitamin D utilization when parents were surveyed) or that the 2 smaller surveys focused on urban populations.
