After choosing a research question, the initial steps of an empirical research project include the following:

• proposing a suitable explanation for the phenomena under study
• formulating testable hypotheses
• defining the concepts identified in the hypotheses

Variables are used to specify how two or more variables are related in an effort to explain the phenomena of interest.

• An independent variable is thought to influence, affect, or cause variation in another variable.
• A dependent variable is thought to depend upon or be caused by variation in an independent variable.
• A variable is a concept with variation, while a constant is a concept without variation.
• In general, more than one independent variable is needed to adequately explain political phenomena.
• A variable that occurs prior to all other variables is referred to as an antecedent variable, while a variable that occurs closer in time to the dependent variable is called an intervening variable.
• An arrow diagram can be used to present and keep track of variables and complicated explanations.
   

When we assert that variation in independent variable X causes variation in dependent variable Y, we are making three assertions:

• that X and Y covary
• that the change in X precedes the change in Y
• that covariation between X and Y is not spurious or a coincidence
   

A hypothesis is an explicit statement about the relationship between phenomena that formalizes the researcher’s informed guess. Data analysis is used to test the hypothesis as it may be correct or incorrect.

There are six characteristics of a good hypothesis. A good hypothesis should:

• be an empirical statement that formalizes educated guesses about phenomena that exist in the political world, not a statement about what the researcher wants to be true.
• explain general phenomena rather than one particular occurrence of the phenomena.
• be plausible—there should be a logical reason for thinking that the hypothesis might be confirmed by the data.
• be specific by stating the direction of the relationship between two phenomena, be it a positive or negative relationship.
• be consistent with the data by using terms that are consistent with the manner of testing.
• be testable through feasible to obtain data that will indicate if the hypothesis is defensible.
   

Hypotheses also specify a unit of analysis, or the level of political actor to which it applies (individuals, groups, states, organizations, etc.).

• Most research uses hypotheses with one unit of analysis.
• While a cross-level analysis specifying more than one unit of analysis is sometimes useful for making ecological inferences about individuals from aggregate data, in general, researchers should not mix units of analysis within a hypothesis.
   

Definitions of concepts should be clear, accurate, precise, and informative, so that others may fully understand the concept as it was tested and evaluate the measurement strategy for the concept.

Many of the concepts used in political science are fairly abstract and require careful and extensive thought to make definitions clear.