110 likes | 305 Views
From Theory to Hypothesis: The Basics of Conceptualization. PO 390: Research Design for Political Science. The “Abstraction” of Theory.
E N D
From Theory to Hypothesis: The Basics of Conceptualization PO 390: Research Design for Political Science
The “Abstraction” of Theory • Thus far, we have reviewed the literature to see what has been done, to identify gaps, and to tailor your research question as a contribution that addresses those gaps or advances discussion • The next goal is to turn the relatively abstract concepts embodied in your theoretical refinement into a concrete proposition that, when tested against empirical phenomena, allows you to gauge the success of your contribution
What Must a Hypothesis Do? • (1)It must refer directly to real-world phenomena while remaining “conceptually relatable” to your theory • The key here is “loose translation” of abstraction into reality • Example: The Diversionary Theory (“Wag the Dog”) • Theory: Domestic difficulties give leaders incentives to pick fights with other states, since such fights lead to “rally effects” that divert popular focus from those difficulties • Concepts: Domestic difficulties, fighting • Phenomena: • Domestic Difficulties – Inflation, Unemployment, Scandal, Low Growth, Poor Approval Ratings • Fighting – Wars, Lower-Level Military Engagements,
What Must a Hypothesis Do? • (2) It must clearly specify some sort of “directional,” causal relationship between those real-world phenomena in a way that also relates to your theory • Hypotheses cannot afford to be vague about relationships; they must constitute a definitive statement of the real-world “footprint” of the causal associations that you have laid out in the theory • Example: “Ceteris paribus [all else being equal], when inflation is relatively high, states are more likely to initiate low-level military engagements abroad than when inflation is relatively low.” • Levels of inflation are clearly related to one another and the dependent variable • The “relation term” (e.g., “more likely”) is sufficiently specific to indicate directionality while sufficiently general to predict for any increase
What Must a Hypothesis Do? • (3) It must clearly establish the temporal, spatial, or other parameters of your expectations • It is possible that your hypothesis, while generalizing about behaviors, is (for any number of reasons) inapplicable to certain places, actors, time periods, etc. • If you have ex ante reasons to believe that the hypothesis should be made more specific, you should tailor it to reflect that expectation (i.e., you should not make unwarranted claims) • Example: Say you have reason to believe one or both of the following: • Only the US has had the military capacity to engage in military conflict “on a whim” • Among the powerful democracies, only the American system of government allows executives to act with the impunity necessary to engage in conflict for domestic political purposes • Thus: “Ceteris paribus, American presidents are more likely to initiate low-level military engagements abroad when facing relatively high inflation than when facing relatively low inflation.”
General “Types” of Hypotheses for Explanatory Research • Conditional: “If presence/absence of X (independent variable), then presence/absence of Y (dependent variable)” • Mathematical: Exact numerical expectation of degree to which one variable is a function of the other (“increase of 1 unit X = decrease of 1 unit Y”) • Continuous: More general (“the greater X, the greater Y”) • Difference: Even more general (“If high X, then low Y”) • Which would you say is best? Why? Can you do it?
The Presentation of the Hypothesis • Two approaches: • “Implicit” hypothesis at the culmination of theoretical and variable specification • Occasionally (as in Foster, 2006) authors end their discussion of theory while leaving the relevant concepts unspecified, and then wait until the delineation of variables to “demonstrate” relationships • While improving the flow of the paper, this method generally presents hurdles to specificity • “Explicit” hypothesis (preferred) • Separate portion of the text • Aim for parsimony, clear language about relationship
The Difficulty of Hypothesis Conceptualization • Process-Tracing: Are there important empirical factors and/or causal relationships missing in your translation? • You must be keenly aware of other empirical factors that might directly amplify, obscure, or otherwise play a key role in your hypothesized relationship. Let’s consider the diversionary example. • Does inflation directly result in diversionary incentives, or does it do so indirectly by lowering public approval? Another words, is the relationship: • Inflation → (+) force, or • Inflation → (-) approval →(+) force? • If you think it’s the latter, which do you specify as your independent variable? Is approval simply a less important “intervening variable” in this equation, or is it what is actually driving diversion? Which relationship(s) should you test? • These questions must be answered in the process of developing hypotheses, and might even require you to further clarify your theory
The Difficulty of Hypothesis Conceptualization • Straw Men: Are you stacking the deck in your favor? • Are you asking a question in almost the precise way that somebody else has asked it? If so, you’re not adding much to the scientific endeavor (repair to the literature!) • Does your relationship seem pretty obvious or uninteresting (e.g., “when the sun comes up, darkness ceases to be”? If so, what’s the value of your research? • Particular danger in developing “negative” hypotheses • Example: “Contrary to previous research, we can’t say that diversionary war happens, because leaders’ll never admit it.” Of course they won’t! But does this mean diversion never happens? Real-world obstacles to generating certain findings must be accounted for in hypothesis construction
The Difficulty of Hypothesis Conceptualization • Over-Specificity: Are you setting yourself up for a fall? • The more specific the hypothesis, the more likely it is to be rejected: does your theory warrant such specificity? • The more specific the hypothesis, the less it explains: how well does the hypothesis meet the key criterion of attaining generalizability?