Much of the work of Mass-Observation, which was founded in England in 1937 by Tom Harrisson and Charles Madge, is concerned with public opinion. A most important problem in opinion survey work is the comparative value of numerical, extensive surveys and more qualitative, intensive surveys. The methods used by Mass-Observation are designed to supplement limited numerical data by qualitative material which assists in the understanding of any figures obtained. This paper attempts to describe the methods used by considering a particular investigation recently completed.