Writing hypertexts: Effects on writing and knowing

1
Writing hypertexts Effects on writing and knowing

• Martine Braaksma, Gert Rijlaarsdam, Tanja
  Janssen, Huub van den Bergh
• Writing Research Across Borders, February 2008,
  Santa Barbara, USA

2
Hypertext writing

• Beneficial effects on
• Writing skills (writing processes and text
  quality)
• Hypertext writing students learn to cope with
  linearization process
• Hypertext writing more planning and analysis
  than linear writing
• More planning analysis better product quality
  of hypertext and linear text
• Content knowledge (topic of writing)
• Hypertext writing more knowledge transforming
  activities

3
Exploratory study

• Aim developing and testing of lesson series and
  testing materials (in cooperation with teachers)
• Design
• Pre-tests
• Experimental lesson series, two conditions
  Hypertext writing (HYP) and Linear writing (LIN)
• Post-tests
• Effects on
• Writing skills (text quality, writing processes)
• Content knowledge
• In relation with learner characteristics
  (aptitude)

4
Design
5
Main overview lesson series

• Five lessons (70 minutes each) on writing
  argumentative texts
• Teacher manual and student workbooks
• All activities in the lessons, no homework
• Theme is good charities, documentation provided
• First 2 lessons based on inquiry learning
  (Hillocks, 1986)
• Two conditions (HYP and LIN)
• First 3,5 lessons exactly the same for the two
  conditions

6
Structure of lessons, main focus
7
Participants

• Two teachers from different schools
• Five 10th-grade classes (senior general secondary
  education)
• Teacher A 1 class HYP 2 classes LIN
• Teacher B 1 class HYP, 1 class LIN
• HYP N 42
• LIN N 63

8
Lesson 1 Making concept maps
9
Lesson 2 Convincing a jury
10
Lesson 3 Making argumentation schemes
11
Lesson 4 Presentation of advices for writing
12
Lesson 5 Writing (hyper)texts
13
Experiences with the lessons

• Positive theme good charities, structured
  organization of lessons and students activities,
  quality of (hyper)texts and other products, pace
  of lesson series, cooperation with peers,
  documentation, writing of hypertexts
  (HYP-condition).
• Advices more explicit instruction for
  composition of structure for hypertexts and for
  taking up standpoint arguments, more feedback
  opportunities (for students teachers), more
  reading time.

14
Some examples of hypertexts
15
Examples (continued)
16
Results Pre-test scores

• No a-priori differences between conditions on
• Computer skills (p .623)
• Aptitude (p .486)

17
Effects on Text Quality (linear text)

• Global Text Quality (school mark between 0-10),
  based on requirements that were presented to the
  students, e.g.,
• Goal of the text
• Attractiveness
• Awareness of the reader

18
Results Effects on global Text Quality
19
Results Global Text Quality

• Regression overall 0.12 (se 0.13), t 0.94
• Aptitude_1 ( HYP-aptitude) regression differs
  significantly 0.26 (se0.13), t1.97.
• So regression HYP-condition 0.12 0.26
  0.38.

20
Effects on Writing Processes (Inputlog)

• Factor Analysis showed three factors (77 of the
  variance explained)
• Laborious Producers (i.e., relative short final
  text, long pauses (not many pauses), during
  session relatively low production, relatively
  long session time).
• Revision Producers (i.e., much produced during
  the session relatively long texts, much text
  deleted during production)
• Balanced Producers (i.e., long session, much
  production and much pausing, featuring many short
  pauses, some deletion during writing)

21
Structure of variable Writing Process
22
Results Effects on Writing Processes

• No differences between conditions on all three
  factors
• No interaction with aptitude

23
Results Writing Process Factor 1
24
Results Writing Process Factor 2
25
Results Writing Process Factor 3
26
Effects on Content Knowledge (concept map)

• Quantitative aspects
• number of concepts and relations
• hierarchy of concepts (number of levels)
• Qualitative aspects
• level of abstraction (categorization of concepts)
  
• variance (different aspects of topic)

27
Example of a concept map
collections
28
Model 1 Differences between measurement
occasions and conditions (Concepts Relations)
29
Model 2 Differences between measurement
occasions and conditions controlled for pre-test
scores (Concepts Relations)
30
Results Effects on Content Knowledge (Model 1
Differences between measurement occasions and
conditions)
31
Results Effects on Content Knowledge (Model 2
Differences between measurement occasions and
conditions controlled for pre-test scores)
32
Reliability

• Aptitude Cronbachs alpha 0.84.
• Inter coder reliability global writing score (2
  coders) 0.915 (Spearmann Brown)
• Inter coder reliability content knowledge (2
  coders) percentage agreement
• Number of concepts and relations 92
• Hierarchy of concepts 95
• Level of abstraction 84
• Variance 89

33
Discussion

• No main effects from hypertext writing on writing
  skills and content knowledge
• Only ATI on text quality HYP
• Possible threats
• Introduction of hypertext writing in the lesson
  series
• Lack of feed back on writing (hyper) texts
• Measurement of Writing Skills (different topic,
  lack of motivation)
• Measurement of Content Knowledge (low observed
  power, no correlations with text quality and
  aptitude)

34
Observed power Content Knowledge (Model 1
Differences between measurement occasions and
conditions)
35
Observed power Content Knowledge (Model 2
Differences between measurement occasions and
conditions controlled for pre-test scores)
36
Future studies

• Study 2 weak and strong learners (N16) follow
  lessons series (HYP or LIN) while thinking aloud
  (Revised lesson series).
• Learning process will be recorded on video
• Insight into learning processes during LIN and
  HYP writing
• Materials as input for study 4 (LIN and HYP via
  observational learning)
• Study 3 effectiveness of HYP versus LIN via
  direct learning (as exploratory study) (Revised
  lesson series).
• Study 4 effectiveness of HYP versus LIN via
  observational learning

37
More information

• Lesson materials, research plan, pictures,
  examples hypertexts, slides at
• http//www.ilo.uva.nl/homepages/martine/hypertext
  _project.htm
• Email M.A.H.Braaksma_at_uva.nl