Anthropological theory and ethnographic data are rich and immediate tools which can be used to create the cultures of fantasy settings. The question is not whether they are useful, but the degree of their utility. Using these methods for cultural construction in a fantasy setting does, however, introduce possible inconsistencies in base assumptions that may conflict with the principles on which anthropology is grounded (being that it is based on the real world). Regardless of complications, anthropological theory is a fundamental framework from which we can begin to ask the right questions and develop a methodology.
Structuralism
The theoretical framework that I have found the most useful is structuralism. In Structuralism, meaning is produced and reproduced within a culture through various practises, phenomena and activities that serve as systems of significance. A structuralist may study activities as diverse as food preparation and serving rituals, religious rites, games, literary and non-literary texts, and other forms of entertainment to discover the deep structures by which meaning is produced and reproduced within a culture. - Wikipedia
Structuralism is very good at providing a generalized framework to work from, based on your initial ideas about fantasy race and culture. The reason I have found structuralism to be the most useful is that it allows you to generate base assumptions suitable to your fantasy setting and then derive "consequences" and/or implications for those assumptions. Basically this sets up an if x then y proposition that serves a designer well. Overall I found that the structuralist framework provided by various anthropologists fit well together with little conflicts and covered all of the major areas that need to be addressed: society, culture, religion, ritual, etc. Structuralism is not, however, without its drawbacks. Quite often the level of abstraction that structuralism provides, in order to provide generalized and/or universal theories can lead to frameworks that are too static. In these cases the problem is typically idealism and over generalization without allowances for the inner dynamics of culture and society. In other words, sometimes the theoretical framework can be too broad and too simple. If we stop there we end up with a very two dimensional culture. If we want something richer we need to proceed to delve into the inner dynamics. In these cases inferences based on creativity are needed to allow further theoretical complexity. These inferences sometimes require changing basic assumptions or properties of our race or culture. Keep in mind that in such an exercise we are working backwards. Anthropologists move from culture to ethnography to theory. Certainly there is something lost in the transition. One example is the statement made by Mary Douglas, "... when the social group grips its members in tight communal bonds, the religion is ritualist; when this grip is relaxed, ritualism declines. And with this shift in forms, a shift in doctrines appears."1 While, not irrelevant it is sufficiently vague as to only give us a general idea of how to proceed.
Ethnography
Once a theoretical framework is sufficiently complex, ethnography provides a rich supply with which to base the socio-cultural details. The problem with applying ethnography to the creation of a culture is that, "as is", it usually isn't really relevant, even the parts that you are looking for that seem to fit. On first inspection it seems easy to just take ethnographic data and use it to fill in your framework, however, this is not a trivial task. Ethnographic details within cultures are the results of complex, interacting symbol systems which have a high degree of multivocality. The symbol systems of a culture are usually fairly integrated with one another and as a result you cannot just remove a symbol system out of context and apply it, without changes, into the symbol system you are trying to create. The underlying symbol systems that sustain and/or gave rise to the system in which you are interested are very often not completely compatible with the underlying symbol systems within your fictitious culture. The best way to approach this task is to look across cultures for those with symbols, themes and basic preconditions, such as environment, physiology, subsistence, and kin structure, that parallel your own. If you have built your theoretical framework correctly then adapting cultural details from a real to a fictitious culture becomes much easier. , but you still need to understand the context of the systems from the culture from which you are borrowing in order to make an accurate translation. This can most easily be accomplished by understanding why a particular culture "does what it does." From that you can, with a little creativity, translate between systems.
Application
The main problem with using anthropological techniques to develop fantasy cultures is that in fantasy the basic preconditions of culture, or combinations thereof aren't always mirrored, or may conflict with, those we find in the real world. When I say basic preconditions I am referring to four specific categories: physiology, environment, subsistence and kin structure (which encompasses gender relations, among other things.) From my observations it is from these four areas that we see rich cultures and cultural patterns emerge. The world of fantasy provides a great deal of flexibility in that you can bend the rules precisely because it isn't real (look at puff the magic dragon; physically it would be impossible for his tiny wings to support his huge body in flight). The danger in this is, of course, bending or breaking too many rules. The best "lie" is the one with the most truth, so we want to stay as close to reality as possible to have good, believable fiction. It isn't hard to see how these preconditions might conflict. For example, I think you would be hard pressed to find an anthropologist that has studied people living in a landscape ravaged by magic or magical storms and alterations. Such a large allowance of the fantastic in such a basic and fundamental area such as the environment makes it difficult to bring the weight anthropological tools to bear on its influences upon the life that exists in said environment. The most we can do in cases such as these is attempt to draw less fantastic parallels and extrapolate the plausible possibilities.
The area in which structuralism is absolutely indispensable is in describing and providing causal relationships between society, culture and the environment. You can do this by moving through the four categories mentioned above in the following order: environment, physiology, subsistence and kin structure (including gender relations.) The natural consequences from one create the basis for the next, and so on. It has been my observation that these naturally flow together, building upon one another. First start with the environment, design it, and then build the next precondition (physiology) based on conditions determined in the environment phase. You may have to do some research to determine how your environmental conditions impact human or animal physiology. Continue this through the last precondition. Each phase proceeds from the previous phase wherein you iterate over its design, test your design against real world examples, adjust as necessary, repeat, until that phase is complete. From these four basic preconditions arises the basis for other systems of society: social structure, religion, ritual, economics and trade, cultural themes, classifications or systems (e.g. Food categories, Leech), history, myths, creation mythology, and more. Once you have a sufficient amount of material for the first four preconditions the rest will, more or less, suggest themselves. The frustrating thing that a designer will notice is that for every question you answer, you seem to end up with a lot more. The amount of work and detail grows on an exponential scale. This is an unavoidable consequence of modelling a realistic cultural system, you also end up modelling the complexity as well. At a certain point, however, you must make a decision of exactly how detailed you wish to get. For your purposes you may be able to gloss over small details, or just provide one or two examples for each (like myths, legends, or heroes). It is important to note that the realism that this approach offers depends on the details. Without the details you end up with an idealized or generalized culture that will not bring your world to life for your readers.
Conclusion
Anthropology and ethnography are invaluable tools for creating cultures in the process of world building, however, there are considerations that an author has to make while bring to bear the weight of these tools on his creativity. Such considerations can leave a writer or researcher walking a fine line between too much fantasy and too much realism. In either respect there is a great deal of work involved in this non-trivial exercise.