Hypotheses
Political Science 5335
Problems in
Patricia M. Shields
Hypothesis --- Statement of expectations. It must be possible to collect evidence that will either support or fail to support the expectation.
Hypotheses are never proven. They are
supported with empirical evidence.
Working Hypothesis --- This is a term borrowed from John Dewey's pragmatism. It deals with expectations that are preliminary in nature. Expectations that are subject to revision. Sometimes working hypotheses are referred to as propositions.
The expectations may be either simple or complex. They may deal with direct facts or relationships between factors.
These are usually the hypotheses used to guide qualitative research such as field or case studies. These are used in formative evaluations.
It still must be possible to collect evidence with can either support or fail to support the hypothesis.
Formal Hypotheses --- These hypotheses generally deal with relationships between concepts or variables.
Loosely the relationship described in an hypothesis is If A then B. This is as close as empirical science can get to a causal link. Or, knowing information about A will tell you something about B. Most empirical studies in academic journals use formal hypotheses.
Formal hypotheses are used in experimental and quasi experimental design. They are used in formal evaluation research. (Program A influences outcome B).
A formal hypothesis can be either
conceptual or operational.
Conceptual hypotheses are at the abstract level and are not tied to a study
design.
Operational or operationalized hypotheses deal with variables (concepts that are
measured). Operational hypotheses are directly linked to data and are
directly linked to the formal study design.
Null Hypothesis A type of formal operational hypothesis stating that two variables are not related in the population. It is the null hypothesis that is actually tested by tests of statistical significance.
Generally a null hypothesis is paired with a formal research hypothesis. In program evaluation the guiding formal hypothesis would be that the program influences the outcome. The null hypothesis would be that there is no relationship between the program and the outcome. If the null hypothesis is rejected the formal hypothesis is supported by the evidence.