Research on intensively farmed animals over the past 25 years has shown that human-animal interactions, by affecting the animal's fear of humans, can markedly limit the productivity and welfare of farm animals. This article begins to explore some of the factors that need to be considered to investigate Keeper-Animal Relationships (KARs) in the zoo. In the mid-1990s, a large body of multi-institutional data on zookeepers and animals was collected from 46 Zoos. Using standardized questionnaires, 82 keepers rated how they behaved towards animals, their husbandry routine, how the animal responds to them and to other people, and provided information about themselves. These data include 219 individuals of four endangered species: black rhinoceros, cheetah, maned wolf, and great hornbill. At each zoo, keepers were also videotaped calling to their animals in order to directly observe animal responses to keeper behaviors. Principle Components Analysis reduced eight animal variables to three components and ten keeper variables to five components. Scores for animals and for keepers were calculated on these components and compared, according to five predictions based on models of human-animal interactions in the literature. Animal responses to keepers varied along three dimensions: Affinity to Keeper, Fear of People, and Sociable/Curious. Animal scores of Fear of People were significantly and positively correlated with independent measures of poor welfare from two later studies: fecal corticoid concentrations for 12 black rhinos and "tense-fearful" scores for 12 cheetahs. (1) Significant species differences were found for Affinity to Keeper and Fear of People, and the interaction of these two dimensions of animal response to keepers appears to be species-specific. (2) The quality of KAR is influenced by whether the zookeeper goes in the enclosure with the animal or not, the frequency and time of feeding, and keeper visibility to the animal. Among keepers who go in with their animals, a significant negative correlation between Frequency of Feeding/Early Feedtime and average Affinity to Keeper of their animals, and a positive correlation between Keeper Experience and their animals' Fear of People, indicates that certain zoo keeping styles or habits among experienced keepers might be aversive and increase fear among animals. (3) Keepers who locomote or make unexpected noises when calling their animals elicit increased aggression or apprehension from maned wolves and cheetahs. (4) Wild-born black rhino and parent-reared maned wolf have significantly less affinity to keepers than their captive-born or hand-reared counterparts, but neither differs in Fear of People. (5) Keeper-animal relationships are likely to be reciprocal as evidenced by a negative correlation of Job Satisfaction with animal Fear of People. Future research directions are discussed with respect to assessment of keeper attitudes and behaviors, animal fear, positive measures of welfare, and positive reinforcement training.