The purpose of an analysis is to bring about the artful disappointment of the reader’s expectations of your object. Less dramatically, we can say that there should be an interesting tension between the theory and the analysis presented in a paper. This is why it is so important to gain the reader’s trust in your methods section. If your readers are more likely to reject your results than let them challenge their expectations then you will not be able to write a compelling analysis. You will only be able to confirm their preconceived notions about reality.
The analysis section will often be the longest and most substantial part of the paper. As a rule of thumb, make it three times longer than your methods or theory section (whichever is longer) or twice as long as the two combined. (This is a very, very rough rule of thumb. Do whatever suits you and your readers, of course.) It will normally be divided into subsections by hypothesis or theme.
In each paragraph, make your key sentence about the object and support that claim with your data. Since it is about the objectified practice you are studying, it will often be expressed in theoretical terms. The meaning of the key sentence, we might say, will depend on your theory, but its truth will depend on your data. Always remember that you are not merely reporting on your data (though you are of course presenting your data); you are reporting what the data shows us about the object. You have earned the right to talk about the object (the real-world social practice you are interested in) in your theory and methods sections. As long as you use the terms you have defined within the limits you have specified, all is well. Don’t limit yourself to talking about the sample. Talk about the population you have sampled from.
For example, your key sentence in a particular paragraph may be something like, “The employees are generally dissatisfied with the food in the cafeteria.” But in the rest of the paragraph you will talk about your basis for saying this; you will draw, for example, on the transcribed semi-structured interviews that we were told about in your methods:
Jake said that he tried to eat there as rarely as he could and stuck to the salad bar when he had no choice. ‘Never eat the soup of the day,’ said Janet, who suspected that it consisted mainly of leftovers from the day before. ‘I think I speak for everyone,’ said Harriet, ‘when I say that the food sucks.’ None of our interviewees said they liked eating in the cafeteria and several offered their distaste for it spontaneously in the context of discussions of other issues.
Notice that we pass from a general judgment about employee “satisfaction” (presumably a theorized notion) to the evidence that our interviews provide. Also notice that the absence of information in the data set can sometimes be relevant, e.g., when supporting a generalization.
When writing your analysis speak plainly and assertively, with the confidence that the rigor of your theory and methods affords. While it can be quite implicit in any particular paragraph, try to be mindful of the disappointment that the claim constitutes for someone who holds the expectations suggested by your theory. If you’re going to say there is general dissatisfaction with the food in the cafeteria, do try to pitch it into the reader’s mind at a challenging angle. You don’t have to make it explicit, but be aware of the change you are trying to bring about. What is your reader supposed to learn from this particular result?
Obviously, the analysis should culminate in a general judgement about the people you have studied. You might conclude, for example, that “XYZ Corp has as strong organizational culture, grounded in a technical rather than social ethic.” (This might be framed theoretically–i.e., in the theory section–by the expectation that strong organizational cultures require a social ethic, fostered by, e.g., frills like good food in the cafeteria.) Make sure that all the individual claims you make, paragraph for paragraph add up this larger judgment. Some will provide evidence of a strong culture. Some will provide evidence of a technical ethic. Some will provide evidence of the absence of a social ethic. Etc.
Don’t get bogged down in the problem of “presenting your empirical material”. This often appears to be an overwhelming task. You will often have much more material than you need; certainly, you will have more than you can present in a single paper. Instead, think of your data as a resource you can use to support the claims you want to make. Carpenters don’t try to use all the wood in the woodshed to make a particular table. They don’t see the variety of materials they have available as a problem to be solved. It contains the means of the solution to the real problem. Think of the abundance of your data in the same way. It’s a good thing.