Establish facts . . . to find whether hypotheses are true or not by eliminating alternative hypotheses

Understand the meaning of phenomena . . . to construct and vindicate descriptions of the world that are deeper, richer, and more useful

positivism
empiricism
operationism
reductionism
objectivism
atomism
mechanism

phenomenology
holism
hermeneutics
constructionism

Science is value-free.

Models come from synthesis of raw empirical data into theories.

Adherence to the scientific method guarantees that results will be context-free and universally applicable.

Paradigms, values, and models guide:
a) questions asked
b) methods used
c) possible explanations
d) dissemination of findings

Scientific research is always influenced by its context. There are no purely "raw" data.

Usually large number of "subjects."

Researcher stays distant from subjects to not introduce bias. Contact may even be automated.

Usually small number of "informants."

Researcher likely to be involved with informants in a relationship based on trust and empathy.

Quantitative

Qualitative data may be used to generate hypotheses which are then tested in an experiment

Quantitative and/or qualitative

Quantitative data may generate questions to be explored qualitatively or vice versa

Experiments
Quasi-experiments
Correlations
Naturalistic observation
Survey research
Meta-analysis

Qualitative methods such as
In-depth interviews
Naturalistic Inquiry
Ethnography
Participant observation
Hermeneutic analysis
Conceptual analysis
AND Quantitative methods

Experimental method is ideal. Other methods are seen as more-or-less deficient approximations of true experiments. They are used only if an experiment is not possible.

Method is selected before data collection and not modified.

There is no "recipe" or research that, if followed, will guarntee truth. Each method has value depending on the question. A combination of methods is desirable.

Method and research questions are integrated and evolve together. Both methods and questions may change during the course of the research.

Control is central

Goal of an experiment is to eliminate chance, individual differences, and other sources of error by holding all extraneous variables constant (i.e., controlling them). This provides evidence that the Independent Variable is the cause of the variations in the Dependent Variable.

Replication is crucial.

Context is central

Focus is on the gestalt of a phenomenon, not isolated variables. Since all research is context-bound, replication and control are illusions. Individual differences, uniqueness, and novelty are crucial parts of human experience and are to be valued in research.

Five dimensions of rigor can be recognized. Both approaches include all five, but in different ways. (See Lincoln & Guba, 1985)

Clarity

operational definitions

conceptual analysis

Truth-value

internal validity

credibility

Applicability

external validity

fittingness

Consistency

reliability

auditability

Neutrality

objectivity

confirmability