Katseline semantika: planeerimine ja teostus
EXPERIMENTAL SEMANTICS: DESIGN AND IMPLEMENTATION

Summary/Abstract: In semantics, conducting corpus linguistic analyses is much more common than conducting experiments. There are various reasons for this bias in empirical methods, including, among other things, the conviction that conducting an experiment is very difficult and risky. The aim of the present article is to discuss some of the practical aspects involved in experimental semantics. The focus is on two crucial stages; planning and conducting linguistic experiments. The article is concerned, inter alia, with the following topics: postulating and operationalising hypotheses, the validity and reliability of an empirical quantitative study, different types of variables, the representativeness of the sample, simple random sample, sample size, choosing an experimental design, constructing and presenting experimental materials, recruiting and instructing participants, and ethical considerations in experimental research. The aim here is not to give a complete list of all the possible practicalities one may encounter when conducting linguistic experiments. The specific design and implementation of an experiment depends largely on the specific research question. The focus is instead on some of the basic aspects that every researcher needs to consider when doing experimental semantics. One of the crucial points in designing any empirical study is the postulation and operationalisation of hypotheses and variables. Taking into account the complex nature of semantic studies, this aspect requires both creativeness as well as vast experience. At the same time, the quality of an experiment (its validity and representativeness) depends largely on the decisions the researcher makes in these early stages of research planning. Although it may seem tedious to spend too much energy and resources on detailed planning, it is worth paying meticulous attention to the crucial step of experimental design. Neither large quantities of data nor advanced statistical methods will save a linguistic experiment that suffers from dubious validity.

• Details
• Contents