Don Carson’s review of Chris VanLandingham’s Judgment and Justification in Early Judaism and the Apostle Paul includes sound advice about two ways to approach writing a dissertation (Carson’s advice also applies to writing a research paper):
I frequently tell my doctoral students as they embark on their research that dissertations in the broad field of the arts disciplines, including biblical and theological disciplines, can, at the risk of slight oversimplification, be divided into two camps.
[1. Deductive Approach] In the first camp, the student begins with an idea, a fresh insight, a thesis he or she would like to test against the evidence.
[2. Inductive Approach] In the second, the student has no thesis to begin with but would like to explore the evidence in a certain domain to see exactly what is going on in a group of texts and admits to uncertainty about what the outcome will be.
 The advantage of the first kind of thesis is that the work is exciting from the beginning and directed by the thesis that is being tested; the danger is that, unless the student takes extraordinary precautions and proves to be remarkably self-critical, the temptation to domesticate the evidence in order to defend the thesis becomes well-nigh irresistible.
 The advantage of the second kind of thesis is that it is likely to produce more even-handed results than the first, since the researcher has no axe to grind and is therefore more likely to follow the evidence wherever it leads; the danger is that there may not be much of a thesis at the end of the process, but merely a lot of well-organized data.
In reality, of course, dissertation projects regularly straddle both camps in various ways. But VanLandingham’s work neatly falls pretty exclusively into the first camp. That makes for interesting reading. Unfortunately, VanLandingham’s work also demonstrates in a superlative fashion the dangers of this sort of approach.