|
| |
| |
| |
Abstracts Volume 28 no. 3 2006
SPECIAL ISSUE: Luuk van Waes, Mariëlle Leijten and Thea van der Geest (guest editors). Methods and Instruments for Reading and Writing Research
| |
Thea van der Geest, Marielle Leijten and Luuk van Waes
University of Twente, The Netherlands / University of Antwerp, Belgium
Researching language production and processing with the computer
ABSTRACT: This special issue discusses newly developed applications of ICT which can be used in writing and reading research. These high tech instruments and methods open new perspectives for language researchers. The various articles describe different research techniques: thinking aloud, keystroke logging, eye tracking and neuroimaging. The ICT applications enable researchers to record, visualize and analyze language data in great detail. The possibilities and limitations of these methods and instruments are discussed and illustrated with exemplary studies. Special attention is given to the methodological issues of reactivity and triangulation.
KEYWORDS: research methods, research instruments, language production, language processing, registration methods, observation methods, visualization methods, reading research, writing research, reactivity
| |
Maaike van den Haak, Menno de Jong and Peter Jan Schellens
University of Twente / Radboud University Nijmegen, The Netherlands
Think-aloud protocols and usability testing: Completeness and reactivity of concurrent think-aloud protocols
ABSTRACT: The think-aloud method has evolved into a widely used tool for a range of studies of reading and writing processes. This paper focuses on the validity of the thinkaloud method as a formative evaluation tool. The validity of the method is discussed in terms of reactivity and the degree of completeness of the verbalizations. The paper explores to what extent the nature of the test object may affect reactivity and completeness. To this end, concurrent and retrospective think-aloud protocols were compared for two types of internet applications: an online library catalogue and a municipal web site. Results showed that for both applications the concurrent think-aloud protocols revealed fewer verbalized problems than the retrospective think-aloud protocols. The nature of the test objects did indeed affect the reactivity of the method: the instruction to think aloud while working with the test object led to more observable problems in the task performance of the library catalogue, but not in the task performance of the municipal web site. This result can be explained by the | |
| |
fact that the two test objects involve different kinds of task performance.
KEYWORDS: concurrent and retrospective think-aloud protocols, validity, reactivity, formative evaluation.
| |
Luuk Van Waes and Mariëlle Leijten
University of Antwerp, Belgium
Observing writing processes with inputlog: A data analysis of the interaction with the text produced so far
ABSTRACT: The use of computers as writing instruments has not only had a profound effect on the writing practice and the attitudes towards writing, it has also created new possibilities for writing research. In this paper we describe Inputlog, a Windows-based logging program. It enables researchers (i) to collect writing process data from different input sources (keyboard, mouse and speech), (ii) to analyze these data from different perspectives (product, process, pausing behavior) and (iii) to integrate them with data logged in other programs. A play module (iv) offers the possibility to review a writing session. In this article we describe the main technical and functional characteristics of Inputlog. To illustrate its possibilities we describe the data from an experiment in which we asked participants to detect and correct discrepancies between the ‘text produced so far’ (TPSF) and their mental representation of what the text should be. Text production is shown to be influenced most by error span - whether the error spans less or more than two characters - and by mode of writing - whether text has been generated by speech recognition or keyboard. This research also shows how working memory load during error analysis affects strategy choice and success as measured by a series of online measures of text production. We conclude the paper with a preview of the plans for further developments (http://www.inputlog.net).
KEYWORDS: Inputlog, keystroke logging, registration tool, on-line writing processes, writing modes, pause analysis, cognitive processes, error analysis, speech recognition, text produced so far (TPSF), working memory.
| |
Reinier Cozijn
Tilburg University, The Netherlands
Analyzing eye movements in reading research
ABSTRACT: The use of eye movements in reading research is explained in a case study on the influence of bold type font and bullet lists on the comprehension of short explanatory texts. The article focuses on how meaningful reading times can be calculated from eye movements and discusses the computer program Fixation which allows for easy analysis of eye movements and data preparation for statistical testing.
KEYWORDS: reading, reading research, eye movements, reading times, measures, software, fixation
| |
Leo Lentz, Pim Mak and Henk Pander Maat
Utrecht University, The Netherlands
Using eye tracking in usability research
ABSTRACT: Does eye tracking help us to detect problems readers experience with documents? In an experimental setting, respondents read an information leaflet about a financial product. Two independent | |
| |
variables were used: participants read the leaflet either silently or aloud, and they either did or did not give comments on the document after reading the text. All respondents did a performance test on the comprehension of ten lexical items. The results showed that reading-aloud and the assignment to comment on the document both have consequences for the reading process. The technique used in the experiment did not allow us to detect lexical problems.
KEYWORDS: eye tracking, web usability, think-aloud protocols, text comprehension, comprehension of lexical items
| |
John Hoeks, Petra Hendriks & Gisela Redeker
University of Groningen, the Netherlands
Communication and the brain: using neuroimaging to investigate dialogues
Speakers and listeners involved in a conversation seem to adhere to specific pragmatic assumptions (Grice, 1975). Speakers, for instance, are supposed to say as much as is necessary, not more, but also not less. In the current study, we investigate how listeners (or better: their brains) react when this pragmatic rule is violated. This sort of question can be answered by making use of what are called ‘neuroimaging’ methods. By means of neuroimaging the workings of the active brain can be visualized. We present a critical review of the most commonly used neuroimaging methods (PET, fMRI, and ERPs), and we describe an experiment to illustrate the use of ERP registration. In that experiment, participants were presented with written question-answer pairs in which the Gricean Maxim of Quantity was violated. This violation was found to be associated with a specific ERP component that, to our knowledge, has not been found before. In addition, other effects were found that point to thematic and syntactic reanalysis of the critical sentence. We conclude that the violation of the Gricean Maxim of Quantity is signaled immediately, and leads to considerable processing difficulty.
KEYWORDS: psycholinguistics, pragmatics, neuroimaging, PET, fMRI, ERPs
|
|
Auteursrechtelijk beschermd