Answer by Eric Miller:
In some cases, it is not. Some anthropologists embrace it as one of the humanities, like history.
On the other hand, anthropologists can and do engage in quantitative social research, using statistical methods used in other fields. As such, it is as scientific as sociology or other social science. As with other questions in science, these are used to test hypotheses.
Further, even ethnography, participant observation, and other qualitative research methods have scientific value. While field observations, unstructured interviews, and the use of inductive reasoning to come to conclusions is a very different method than experimentation and surveys, they can be part of the scientific endeavor human behavior. These methods help us to decide which hypotheses to test, and to ask the right questions in survey instruments.
From the perspective of sciences like physics, the study of human behavior is necessarily complex. Even after carefully testing hypotheses, the true meaning of the results is often unclear. If neat and tidy answers is a hallmark of science, then anthropology, and indeed any research on human behavior is surely beyond the realm of science (but I do not believe this is the case.)
Qualitative methods can be used to further explore the meaning of results. Because the subject of study is humans, we have the ability not only to ask them questions, but also to ask them to clarify their responses. From there, we can develop the next set of hypotheses and further investigate our study of the subject in question.
Any study of human behavior and society which fails to account for the specific context and its cultural and historical background will fail to correctly interpret the data. (This includes DNA research which at first appears to be immune to such influences, but in reality is not.) As a result, cultural anthropology provides a critical perspective for correct interpretation of all scientific data relating to human behavior.