How to research
Some Observations on Variables, Relationships, and Hypotheses
The empirical approach
•“Empirical” means grounded in experience or observation
•As distinguished from deductive theory
•Requires a RW referent of some sort
•Observation interacts with theory
•Good empirical research has implications for both theory and practice
Characteristics of successful research
•Activity and personal interest of the researcher
•Convergence of interests
•Real world focus
Characteristics of not-so-successful research
•Lack of theory
Defining problems for research
•Why are we interested?
•What are the relationships between our concepts that we hope to find?
•What are the possibilities for finding out about it?
•How are we sure we’ve found anything meaningful?
Why are we interested?
•The phenomena are definable
•The phenomena might be changed
•The consequences of the phenomena are real
•Settings are available where we might look at the phenomena
•Resources are available for the inquiry
What are the expected relationships?
•What does general theory say?
•What do other researchers claim to have found?
–Do we believe them?
•What do we personally expect to be happening?
Forming good questions
•Who needs the information?
•What decisions will be based on it?
•What affects the decisions?
•Who are you asking?
•What are the consequences of wrong answers?
The “efficiency principles”
•Bound the problem carefully
•Pick the simplest design you can given your tradeoffs
•Use labor intensive methods only when really necessary
•If it’s not worth doing, it’s not worth doing well
Research process tradeoffs
•In Research Writing, most common approach
•Usually varying emphasis
•Define the domain of a study
•May vary in specificity
•Should be limited in number
•Should enable the formulation of hypotheses
Levels of analysis
•Measurable properties of things that we choose to notice
•Selected from a potentially infinite list
•Come in several “flavors”
–Expressed in numbers (“quantitative”)
–Expressed in words (“qualitative”)
•Are statements of expected relationships
•Uses operational measures
•Include 2 or 3 variables
•May be defined over large or small groups
•May be multiple
•Should be directional
•Implicitly includes its “null”
Mapping a relationship
•Dependent variable right, independent left
•One ways arrows indicate hypothesized causality
•Valence signs indicate direction of relationships
•Two-headed arrows indicate unanalyzed correlations