Exploratory factor analysis is a form of factor analysis that emphasizes "exploring" a data set for relationships among the variables. What a researcher finds through an exploratory factor analysis can assist the researcher in developing a model or a set of hypotheses regarding the variables of interest in the data. An important aspect of exploratory factor analysis is its ability to take large sets of variables and reduce them to a set of new variables, known as "factors," which can play the role of a new, smaller set of variables for further study.
The basic procedure of exploratory factor analysis is to enter a set of data, run the common factor model (a model that allows variables to combine themselves and form factors) and rotate the solution so it is easily interpretable. The part of this procedure that is of most interest is the rotation, which is the most important job in factor extraction, in which the researcher rotates the solution to yield a logical result.
The role of factor extraction in factor analysis is to take the set of variables originally entered with the data and place them on a set of "extractable" factors. Factor extraction includes the common factor model solution and the rotation of the solution. However, only the rotation of the solution is alterable, so this is what researchers refer to as the "technique" of factor extraction.
After running the common factor model in factor analysis, a researcher will have a set of factors. However, these factors are often not easily interpretable. The researcher then performs techniques of factor extraction, which are methods of rotating the solution so that the final set of factors makes logical sense and can be re-related to the original variables. The two main techniques of factor extraction are oblique rotations, which allow correlation between factors, and orthogonal rotations, which maintain zero correlation between factors.