Information Technology Reference
In-Depth Information
A variable that has been show to have a considerable effect on a reader's com-
prehension is the
coherence
(or cohesion) of written text (Britton & Gulgoz, 1991;
McNamara & Kintsch, 1996; Vidale-Abarca, Martinez, Gilbert, 2000). McNamara
(2001) defines coherence as the extent to which ideas in the text are made explicit.
Two types of text coherence are specified (Ainsworth & Burcham, 2007): global
and local.
Global
coherence involves using (a) topic headings that summarize the
block of text that follows and (b) macro-propositions that link consecutive para-
graphs to each other and to the topic.
Local
coherence is achieved by (a) replacing
pronouns with nouns when the referent is potentially ambiguous, (b) linking unfa-
miliar concepts to familiar concepts or to previous information presented in the
text, and (c) using connectives that specify relations between consecutive sen-
tences (Ainsworth & Burcham, 2007). A variety of dependent measures have been
reported (McNamara, Kintsch, Butler Songer, & Kintsch, 1996; McKowen, Beck,
Sanatra, & Loxterman, 1992), ranging from shallow definitions and explicit knowl-
edge taken directly from the text to implicit questions and inference questions that
involved deeper knowledge and required integration of textual information and the
use of prior knowledge (Kintsch, 1998). While increased coherence has little effect
on surface knowledge of the text, like definitions, deeper knowledge reflected in
implicit questions and inferences increases significantly. Coherent text apparently
helps learners compare an existing mental model with a model presented in the text,
while detecting any flaws or knowledge gaps, and repairing them to bring that model
in line with the one presented in the text. As indicated directly above, methods for
improving text coherence are straightforward and readily implemented (Ainsworth
& Burcham, 2007, p. 291).
Two features of the speaker's voice, the
style
(formal, personal) and the
quality
(accented, non-accented, synthesized) have been shown to affect comprehension
(and learning) in monolog presentations. Mayer, Fennell, Farmer, and Campbell
(2004) explored voice style in three experiments by presenting an animation depict-
ing (a) how air is inhaled into the lungs, (b) how carbon dioxide and oxygen are
exchanged, and (c) how air is exhaled (p. 390). The animation was accompanied
by a 100-word narration in either formal or personalized style. Both narrations
were spoken by a male with a standard unaccented American voice (p. 391). In
the formal style the narrator said, for example, “the diaphragm moves
...
,” “for the
...
...
lungs
.” In the personalized style the word “the” was
replaced 12 times in the passage by “your.” Thus these participants heard “your
diaphragm moves
,” and “through the throat
,” etc. In each of the three studies comprehension was signif-
icantly enhanced by narration presented in personalized style, in that scores were
significantly higher on transfer tasks involving applications of the content. These
and earlier findings (Moreno & Mayer, 2000) were taken to indicate that personal-
ized narratives containing self-referential language promote better comprehension
through deeper processing of the content (Rogers, Kuiper, & Kirker, 1977; Symons
& Johnson, 1997).
Mayer, Sobko, and Mautone (2003) investigated the role of the voice quality
in two experiments in which participants heard monolog presentations of 16 steps
describing how lightning is formed while they simultaneously watched a 2-min
...
