"Seizure-Alerting and -Response Behaviors in Dogs Living with Epileptic Children" reported in the journal Neurology in June 2004 that 9 out of 60 dogs in families with epileptic children were able to anticipate a pN pN seizure, as evidenced by licking, whimpering, or standing next to the child. According to the families, these dogs predicted 80% of the children's seizures, with no "false positives."7
a. Does "false positive" mean predicting a seizure and it does not occur, or not predicting a seizure and it does occur?
b. Keep in mind that the data were reported by the families themselves, via surveys and telephone interviews. Should we suspect 0.80 to be an overestimate, an underestimate, or an accurate estimate of the proportion of the children's seizures that the dogs were able to anticipate?
c. Could an experiment be designed as a future study to gather data on this phenomenon? If so, describe it. If not, explain why not.