RESEARCH METHODS: EXPERIMENTAL
المؤلف:
John Field
المصدر:
Psycholinguistics
الجزء والصفحة:
P249
2025-10-06
335
RESEARCH METHODS: EXPERIMENTAL
The most widely adopted approach in psycholinguistic research is to gather evidence experimentally. An issue is the extent to which a study under experimental conditions can claim to be investigating the natural use of language.
It is important to specify whether a particular research method taps in to language processing on-line (as it is actually occurring) or whether it involves some kind of subsequent decision-making. If the latter is the case, then the experiment is said to be post-perceptual: it measures the outcome of the process (e.g. what the subject reports they heard) rather than the process itself (the act of listening). Experimenters also have to beware of task effects: a task that requires subjects to recall something they read an hour ago involves not just reading skill but memory as well.
A wide range of experimental methods has been devised, many specific to psycholinguistics.
One group involves measuring how long it takes a subject to respond to a stimulus. The advantage of this kind of method is that it taps into on line performance with no decision-making intervening. The simple tasks that are set are often distractions. They enable the experimenter to measure the effects of a particular condition (for example, sentence ambiguity) upon the subject; but the subject is not aware of the true goal of the experiment and can be assumed to be processing the material normally.
Reaction times (also referred to as response times and response latencies) are usually expressed in milliseconds. Examples of reaction time (RT) tasks are:
Phoneme/word monitoring. The subject presses a button whenever they hear a particular phoneme or word in a recorded text. Their response is slower when the phoneme or word coincides with, for example, semantic or syntactic ambiguities in the text. This enables the experimenter to measure the increased demands which these ambiguities make upon working memory.
The naming task. The subject reads words aloud. The time (naming latency) is measured from the presentation of each word to the onset of the voice.
Lexical decision. The subject presses a button whenever they encounter an actual word rather than a non-word. This task can provide evidence of: (a) how readily different types of word are recognised; (b) how strongly other attention-demanding operations interfere with word recognition.
Word spotting. Subjects hear sequences of sounds which may or may not have actual words embedded in them. They press a button when they identify a word. The task is used to investigate how subjects segment connected speech into words.
These methods provide more than just reaction time data. The experimenter can also record how many items were missed or wrongly identified.
An alternative way of measuring reaction times is to present sentences word-by-word or phrase-by-phrase on a monitor. As a subject finishes reading each display, they press a key for the next, and the time between keystrokes is measured. Equipment is also available which tracks the eye movements of a subject reading a text; this enables researchers to measure fixation times and regressions of the reader’s eye.
A similar group of methods involve detection tasks in which subjects have to determine whether a feature is present and/or where it is located. The researcher might record which features remain undetected or the distance between the actual and perceived position of a feature. Examples are:
Click location. Subjects hear sentences with a click in them, and indicate where they believe the click occurs. The click may be perceived in the wrong place because of the influence of syntactic structure.
Phoneme restoration. Subjects hear sentences from which a phoneme has been excised. They often cannot hear the gap.
Other methods test subjects’ interpretation of linguistic input:
Shadowing. Subjects listen to a cassette recording, and repeat what they hear as closely after the speaker as possible. Shadowing provides insights into spoken language perception, particularly through the errors that subjects make.
Gating. A piece of recorded speech is sliced into sections or ‘gates’. Subjects hear the first section and write down what they think they hear. Then they hear the first and second sections together and add to or revise their first impression. They continue to report on chunks of increasing size.
Masking. Subjects hear sentences with background noise or read sentences where the text is partly masked. They report what they understand. A similar task asks subjects to report faint speech in the form of sentences played at the borders of audibility. The data obtained is post-perceptual.
A further set of experimental methods is used to investigate lexical storage and lexical access. It includes:
Word association. A traditional task for investigating how words are stored in semantic networks. Subjects are given a word and asked to record or write down the first word that comes into their heads.
‘Tip of the Tongue’. The experimenter gives a definition of an obscure word and asks the subject to name the word. This gives rise to many nearly accurate responses, which are compared with the correct one, so as to provide insights into what happens when we search for words.
Priming. A word is presented to a subject, followed shortly afterwards by another believed to be associated with it. If there is a lexical association, the second word will be detected more rapidly. Priming experiments often incorporate the lexical decision task (see above).
Methods used to investigate working memory include:
Recall. Subjects remember as many words as they can from a list.
Repetition. Subjects repeat digits, words or non-words after the experimenter. This forms the basis of various span tests, in which the number of items for repetition becomes longer and longer until the subject reaches a limit, a measure of their memory capacity.
An ingenious set of methods has been designed to deal with the difficulty of acquiring information from pre-linguistic infants. The techniques exploit the fact that infants quickly become bored when they are presented with stimuli that are repetitive. They include the high amplitude sucking (HAS) procedure and the operant head turn procedure. See language acquisition: research methods.
See also: Gating, Priming effect, Shadowing, Verbal report
Further reading: Gernsbacher (1994: Chaps 1–2); Grosjean and Frauenfelder (1996); Jusczyk (1997: Appendix)
الاكثر قراءة في Linguistics fields
اخر الاخبار
اخبار العتبة العباسية المقدسة
