Foreign Language Reading through Hypermedia: Predictors of Vocabulary Learning and Reading Comprehension

1
Foreign Language Reading through Hypermedia
Predictors of Vocabulary Learning and Reading
Comprehension

• Yavuz Akbulut
• Anadolu University, Faculty of Education,
  Department of Computer Education and
  Instructional Technologies
• Eskisehir, Turkey
• yavuzakbulut_at_anadolu.edu.tr

2
Hypermedia

• Computer-based applications that provide
  information in a nonlinear way through multiple
  types of resources such as text, graphics, sound,
  video, and animation (Kommers et al., 1996).

• Notable numbers of software developed for
  language teaching, since hypermedia
• present input in multiple forms
• provide more interaction
• use authentic material more efficiently

3

• Reading Glosses / Annotations
• A glossary is defined as a list of words and
  phrases, and their meanings, which are judged to
  be outside the learner's current competence
  (Widdowson, 1984).
• Glossaries
• help learners to cope with text comprehension
• sustain authenticity better than text
  simplification
• increase
• flow of reading
• independence from dictionaries
• accuracy in finding meaning

4

• Vocabulary Learning
• Vocabulary learning outcomes vary according to
• linguistic proficiency
• individual learning styles
• annotation type

5

• Reading Comprehension
• Hypermedia facilitates reading comprehension
• Reading comprehension outcomes vary according to
• Foreign language reading proficiency
  
• Prior knowledge on the subject domain
• Topic interest
• Learning styles

6
Research Questions

• What are the predictors of vocabulary learning in
  a hypermedia environment for advanced language
  learners?
• What are the predictors of reading comprehension
  in hypermedia environments for advanced language
  learners?

7
Methods Procedures

• Dependent Variables
• Vocabulary scores
• Reading comprehension scores

• Independent Variables
• Annotation type with three levels (i.e.,
  definition, definition plus picture and
  definition plus video), reading ability,
  cognitive and perceptual learning styles, prior
  topical knowledge, topic interest.

• Participants
• 69 freshman students studying at a TEFL
  department in a Turkish university, 47 female and
  22 males.
• They were randomly assigned to three pools and
  those three pools were randomly assigned to
  levels of the treatment.

8

• Treatment Hypermedia Reading Software
• An authentic reading text was selected via a
  topic interest questionnaire. The text consisted
  of 1330 words
• 42 words underlined as unknown by the pilot group
  were chosen as annotations.
• Three forms of the very same text were prepared
  using a reading software designed by Ariew
  (1999). Each form has the same layout and had 9
  pages with a linear organization
• Form 1 Definition only
• Form 2 Definition and associated picture
• Form 3 Definition and associated video
• A tracking tool built in the software was used to
  save every single interaction of the readers with
  the reading material.

9
Instruments (1/3) Vocabulary Test - Similar
tests were used before (Knight, 1994 Rott, 1999
Waring Takaki, 2003) - Consisted of three
parts Form recognition, meaning recognition and
meaning production. - Two pilot studies were
conducted to sustain ideal item facility, item
discrimination and distractor efficiency indexes,
to detect unknown distractors and to determine
ideal timing. - Cronbachs alpha .76 in
pre-test, .72 in post-test, .69 in delayed test -
Interrater R .94 in pre-test, .98 in post-test,
.98 in delayed test
10
Instruments (2/3) Reading Comprehension Test -
Test rubrics were prepared according to Alderson
(2000), Bachman and Palmer (1996) and Urquhart
and Weir (1998) - 12 multiple choice, 22
true-false questions - Proofread by an American
colleague for unidiomatic language - Cronbachs
alpha .79 (multiple-choice), .70 (true-false)
11

• Instruments (3/3)
• Nelson Denny Reading Test (Brown et al.,1993)
• Prior Knowledge Test
• Consisted of a recall part and open-ended
  questions
• Interrater reliability coefficients were.97 on
  the recall part and .96 on open-ended questions
• Learning Style Questionnaire (Oxford, 1993)
• Topic interest questionnaire (Schiefele Krapp,
  1996)

12

• Procedures
• Prior knowledge, topic interest and vocabulary
  pre-test
• - 2 weeks before the treatment
• Nelson Denny Reading Test
• - 11 days before the treatment
  
• Treatment
• - given in a computer lab designed for
  simultaneous processing of 25 computers each
  connected to a local area network (LAN).
• - 3 sessions were arranged for 69
  students, 23 in each.
• - Through a data projector, researchers
  oriented students with the material before each
  section.
• Vocabulary Post-test and Reading Comprehension
  Test
• - Immediately after the treatment
• Vocabulary Delayed Post-test
• - 3 weeks after the treatment

13

• Data Analysis
• Gained vocabulary scores were calculated
• All variables were put into a bivariate
  correlation calculation in SPSS.
• Variables that had high correlations with the
  outcome variables were put into a hierarchical
  multiple regression analysis.
• Variables that had high correlations with each
  other were eliminated from the analysis and
  final regression analyses were conducted.
• Assumptions of observation independence,
  linearity, multicollinearity, non-zero variance,
  normally distributed errors and uncorrelated
  residual term were sustained.

14
Results (1/3)

• Immediate Post-test Scores
• Predictors
• Language proficiency (r.323, plt.007)
• Prior topical knowledge (r.355, plt.003)
• Annotation type (dummy variable)

Predictor Explained Variance (R square) F Change Sig. Beta
Language Proficiency .10 7.801 .007 .258
Prior knowledge .16 7.076 .002 .241
Annotation type .05 4.309 .042 .228
p lt .05 , plt.01
15
Results (2/3)

• Delayed Post-test Scores
• Predictors
• Language proficiency (r.319, plt.008)
• Annotation type (dummy variable)

Predictor Explained Variance (R square) F Change Sig. Beta
Language Proficiency .10 7.590 .008 .306
Annotation type .12 4.828 .011 .163
p lt .05 , plt.01
16
Results (3/3)

• Reading Comprehension
• Predictors
• Reading ability (r.386, plt.001)
• SAS visual score (r.248, plt.04)

Predictor Explained Variance (R square) F Change Sig. Beta
Language Proficiency .15 11.760 .001 .383
Visual score .06 4.887 .03 .242
p lt .05 , plt.01
17
Discussion

• It is better to present an explanation in words
  and visuals than solely in words (Generative
  Theory, dual channels assumption)
• Linguistic proficiency (Knight, 1994) and
  annotation type (Seghayer, 2001) are important
  variables contributing to vocabulary learning.
• Reading ability is the basic predictor of text
  comprehension (Devine, 1988)
• There is a relationship between individual
  learning styles and learning outcomes (Andris,
  1996 Plass et al. 1998 2003)
• Manipulating the instructional design has a
  beneficial outcome for easy materials as opposed
  to Sweller (1994) and Sweller et al.(1998)

18
Pedagogical Implications

• Interaction between the reader and the text
  provides individualized learning.
• Learners can have control over their learning
  process and learn at their own pace.
• Learners with different learning styles can make
  use of hypermedia environments
• Presentation of authentic input is made easier
  with hypermedia software.
• Familiarizing L2 learners with hypermedia
  experience and training them can be quite
  feasible.
• Professionals involved in material development
  should consider the importance of visual aids in
  language teaching through hypermedia.

19
Limitations and Suggestions for Further Research

• The target population of the study was EFL
  students who learn English for academic purposes.
  This study should be replicated in other learning
  contexts.
• Further studies with larger sample sizes must be
  conducted to investigate whether a really
  significant variation existed in the population.
• The use of a pre-test affected the
  generalizability of this study the results
  cannot be generalized to people who are not
  pre-tested.
• Participants were exposed to a non-traditional
  treatment, but tested with traditional testing
  methods in the current study. Employing on-screen
  tests where visual elements are incorporated
  would be more suitable for use in hypermedia
  environments.

20
(No Transcript)
21
(No Transcript)
22
(No Transcript)
23
(No Transcript)