Experiments:

  • enable linguists to investigate in an artificial situation how subjects react when specific details (variables) are changed while everything else is kept constant
  • results are clearly focused and easy to process and evaluate
  • typically involve groups consisting of similar subjects (assigned randomly or shared equally according to relevant factors (gender, age, education, ...))
  • require a statistically viable number of subjects
  • materials: recording equipment, flashcards, a stopwatch, etc.; in many cases computers can do the job: randomising stimuli, displaying stimuli for timed periods, record subjects' reactions to them (e.g. Inquisit: http://www.millisecond.com/)

Possible experimental setups:

  • two identical groups doing different tasks
  • two identical groups doing the same task in different conditions
  • two groups that differ in some specific way (e.g. age, gender, native/non-native speakers)
  • one group doing two different tasks
  • one group doing the same task in different conditions

Examples of linguistic experiments:

Considerations in connection with experiments:

  • inform subjects about purpose of experiment?
  • difficulties in producing unambiguous, comparable stimuli
  • order effects
  • floor and ceiling effects
  • practice and fatigue effects
  • generalizability of results limited because situation is pared down to the bare minimum
  • human subjects involve substantial unexplained variation
  • subjects may feel inhibited; children may have too short an attention span
source: Wray & Bloomer (2006: chapter 12)
Zuletzt geändert: Donnerstag, 2. Oktober 2014, 14:14