Qualitative research involves investigations of concepts or phenomena that are not numerically measurable, such as perceptions or opinions and the reasons behind them. Methods used to gather such information in ESL research include open-ended interviews (with ESL students or teachers) in which both the interviewer and subject feel free to expand upon, diverge from or skip certain questions. Other qualitative methods include classroom observation and analysis of unscripted language use (such as recordings of casual conversations or teacher-student conferences).
Qualitative methods are commonly used in ESL research because they are able to capture the holistic nature of second-language learning. ESL learning is strongly affected by extralinguistic factors such as students' cultural backgrounds and attitudes, so learning can't be accurately analyzed just by looking at language use in isolation. Research methods such as open-ended interviews and classroom observation may also be more agreeable and less intimidating to subjects being studied than participating in experiments or answering written surveys. A potential disadvantage is the perception among some researchers that qualitative methods are less rigorous and objective than quantitative ones, which causes some qualitative research to be taken less seriously.
Quantitative research methods, which deliver measurable, replicable results, are also used in ESL research. Typical methods include surveys and questionnaires with fixed options for possible answers. Written questionnaires in ESL research, for instance, may consist of multiple-choice questionnaires asking ESL teachers about the teaching techniques they use; or surveys on the student demographics of ESL programs.
Quantitative research methods are useful for delivering concrete numerical data and testable theories. Thus, educational policymakers generally pay more heed to results of quantitative than qualitative research. Quantitative methods, however, can only shed limited light on the reasons behind the numbers, and are thus better for describing patterns than explaining them.