The official website of educator Jack C Richards

Collecting Data for Language Research


Submitted by Anna, Instituto Superior Cristal Dili, Timor-Leste

If we conduct language research… what is the technique and procedure to collect the data?

Dr. Richards responds:

Two different kinds of approaches can be used in conducting research, quantitative and qualitative. Quantitative measurement refers to the measurement of something that can be expressed numerically. Many tests are designed to collect information that can be readily counted and presented in terms of frequencies, rankings, or percentages. Other sources of quantitative information are checklists, surveys, and self-ratings. Quantitative data seek to collect information from a large number of people on specific topics and can generally be analyzed statistically so that certain patterns and tendencies emerge. The information collected can be analyzed fairly simply because subjective decisions are not usually involved. Traditionally, quantitative data are regarded as “rigorous” or conforming to scientific principles of data collection, though the limitations of quantitative information are also recognized; hence the need to complement such information with qualitative information.

Qualitative measurement refers to measurement of something that cannot be expressed numerically and that depends more on subjective judgment or observation. Information obtained from classroom observation, interviews, journals, logs, and case studies is generally qualitative. Qualitative approaches are more holistic and naturalistic than quantitative approaches and seek to collect information in natural settings for language use and on authentic tasks rather than in test situations. They are normally more exploratory and seek to collect a large amount of information from a fairly small number of cases. The information obtained is more difficult to analyze because it is often open-ended and must be coded or interpreted. Qualitative data are sometimes regarded as “soft” or less rigorous than quantitative data, but such information is essential in many stages of program evaluation.