The Effects of Audio Narration on Elearning

1
The Effects of Audio Narration on E-learning

• Alyce M. Dickinson
• Wendy Jaehnig
• Western Michigan University
• FABA/OBM
• January, 2007

2
E-learning Explosion

• In the US, money spent on e-learning increased by
  22 in 2003, even though corporate training
  expenditures remained fixed
• Mullich, 2004
• The global market is expected to grow 27
  compounded annually over the next four years
  International Data
  Group, 2006
• In terms of dollars, in 2003, the e-learning
  market was about 6.5 billion and is predicted to
  increase to more than 21 billion by 2008
• International Data Group, 2006

3
A Bit of Background

• Wendy Jaehnig conducted this study as her
  dissertation

4
Audio Narration - Definition

• Presentation of spoken words as principal
  instructional content simultaneously with visual
  instructional components (pictures, diagrams,
  animations, etc.)
• Contrasted with textual narration in which words
  are written rather than spoken

5
Audio Narration
Textual Narration
The title of your HTML page is included in the
HEAD section of the document. A title is
recommended for every web page. Typically, the
title will appear in the title bar of the web
page window. The title is often used by search
engines such as Yahoo, and will be used when
people bookmark your web page in their browser.
6
Audio Narration

• Benefits
• Accessible most authoring tools have the
  capacity to incorporate audio (Authorware,
  FrontPage, Flash)
• Popular used by many developers
• Widely held to be beneficial to learning
• Costs
• Expensive
• Time-consuming
• Quality audio may require paying for professional
  voice talent and studio time (re-record when
  changes are made)
• Resource-intensive (space and bandwidth)
• Slow down the presentation of instructional
  material
• Is it worth doing?

7
Literature Review

• Audio-only narration beneficial for instruction
  that
• has instructional content (charts, animations,
  etc.) that cannot be understood without
  accompanying narration, and vice versa

Audio better than textual
8
Literature Review

• Audio-only narration loses that benefit
• If the visual components require a lot of visual
  searching

Audio NO better than textual
9
Literature LimitationsType of Instruction

• A lot instruction does not involve the type of
  instruction where audio narration has augmented
  learning (integrative instruction)
• Only five studies have examined non-integrative
  instruction
• Results were mixed
• Three found no differences
• One found audio narration to be better
• One found textual narration to be better

10
Literature LimitationsQuality of Instruction

• Only 2 of the 5 studies used active instruction
  in which learners were required to make responses
  and received feedback on those responses
• What we might call good instruction
• Neither found a difference between audio and
  textual
• (Koroghlanian Klein, 2000 Kroll, 1974)
• These were the only 2 studies in the entire
  literature based that
• used active instruction

11
Literature Limitations Quality of Instruction,
cont.

• Even though these two studies used active
  instruction, there was still an issue about the
  quality of the instruction - neither assessed the
  adequacy of the instruction
• In one, learner performance on the criterion test
  was no better than chance, suggesting ineffective
  instruction or an inadequate criterion test
  (Kroll, 1974)
• Thus, while the instruction was active in the two
  studies, was it good active instruction?

12
Literature LimitationsReading and IQ Scores

• Three studies have examined whether the effects
  of audio narration are influenced by reading
  ability or IQ scores
• Results are mixed
• One showed an interaction - Barton Dwyer, 1987
• Two did not - Kroll, 1974 Lai, 2000
• Only one used active instruction (Kroll, 1974)
• No interaction
• Results suggested that either the instruction was
  not adequate or the criterion test was
  inappropriate

13
Summary

• No study has examined the effects of audio
  narration on the performance of individuals with
  different reading skills using tested, effective
  programmed instruction

14
Research Question

• What are the relative effects of audio-only,
  textual-only, and audio-textual narration on
  e-learning for individuals who have different
  reading skills?
• using non-integrative instructional
  material

15
Participants and Instruction

• Participants
• 184 university students (60 or 62 per group)
• Basic computer skills
• Scored 20 or less on a pretest of HTML
• Instruction
• Pretested computer-based programmed instruction
  module teaching HTML
• 83 frames
• 23 response requests that were multiple choice,
  short answer and application questions

16
DVs Covariate

• Dependent variables
• Posttest score on a computer-based criterion test
  (same test used to screen Ps)
• 20 composition questions
• Development of a web page
• Instruction completion time
• Covariate
• ACT reading test score
• 35-minute reading comprehension test
• 40 multiple choice questions

17
Pay Procedures

• 5.00 for completing reading test
• 5.00 for completing instruction
• 10 x percent correct on pretest
• 10 x percent correct on posttest

18
Independent Variable

• Type of narration
• Textual-only
• Audio-only
• Audio-textual

19
Textual-Only
20
Audio-Only
(Through headphones) Because HTML is saved as
Text Only or ASCII, any computer can read HTML
regardless of whether it is an IBM or Macintosh,
and regardless of the operating system used. This
feature is called universality.
21
Audio-Textual
(Through headphones) Because HTML is saved as
Text Only or ASCII, any computer can read HTML
regardless of whether it is an IBM or Macintosh,
and regardless of the operating system used. This
feature is called universality.
22
Experimental Design Procedures

• Between-group experimental design
• Session 1
• Participants took the ACT reading test
• Session 2
• Participants completed the pretest, instruction
  and posttest (at their own pace) in a computer lab

23
Results
24
Results ACT Distribution

• Range 7 - 40
• M 25.67
• Few scores below 12, but MC test with 40
  questions and four choices, individuals would be
  expected to get a score of 10 by chance

25
Main Research Questions

• Was there a difference between the three types of
  narration?
• Did people with different ACT scores benefit
  differentially from the types of narration?
• Actually had to answer the second question first,
  because the answer determined the type of
  significance test that was appropriate

26
Data Analysis - Procedures

• For each of the dependent variables
• First determined whether to use parametric or
  non-parametric statistical analyses
• Then determined whether there was an interaction
  between ACT scores and the type of narration
• i.e., whether individuals with different ACT
  scores benefited differentially from the three
  types of narration
• Finally, determined whether there was a
  difference between the three types of narration
  overall
• Using the significance test that was determined
  to be appropriate in the second step

27
Parametric vs. Non-parametric Test

• Neither dependent variable met the normal
  distribution assumption for parametric analyses
• Posttest scores were negatively skewed
• Instruction times were positively skewed
• Used non-parametric statistical procedures
• Procedures based on Robust General Linear Model
  (RGLM, McKean Vidmar, 1994)
• Rank-based procedures
• Less sensitive to extreme scores
• Used RGLM procedures (a) to determine whether
  there was an interaction between ACT scores and
  type of narration and (b) the ANCOVAs

28
Interaction Between ACT and Type of Narration for
Each DV

• Regressed ACT scores on DVs for each group
• Compared regression slopes for groups
• Homogeneous slopes
• No interaction between IV type and
  covariate/attribute
• ANCOVA to check differences between groups
• Heterogeneous slopes
• Interaction between IV type and
  covariate/attribute
• Johnson-Neyman to determinebest IV for people
  with differentcovariate scores

29
(No Transcript)
30
Results
No difference between the three narration groups
- slopes were homogeneous
31
? .464 p-value 0
32
Results
No difference between the three narration groups
- slopes were homogeneous
33
Results Homogeneity

• People with different reading skills did not
  benefit differentially from audio, textual, or
  audio-textual narration
• ANCOVA was an appropriate statistical
  significance test to test for differences between
  the groups for both posttest scores and
  instruction time

34
Results
35
Results
36
Results
37
Results
38
Summary of Results

• Audio narration had no benefit in this study (but
  neither was it detrimental)
• Audio-only, Textual-only, and Audio-Textual
  narration did not differ with respect to their
  effects on posttest scores or instruction time
• Learners with different reading skills (as
  measured by ACT reading test) did not benefit
  differentially from the three types of narration

39
Discussion

• Practical implications Two considerations when
  deciding what type of narration to use
• Does the instructional content contain narrative
  and visual components that cannot be understood
  alone, and is little visual searching required by
  the visual components?
• YES Audio-only narration
• Is marketability a major concern?
• YES Audio-textual narration
• If NO to both Textual-only narration

40
Future Research

• Replicate with instruction that has different
  subject matter characteristics
• i.e., more emphasis on teaching vocabulary
• Replicate with different populations
• i.e., poorer or better readers
• M of 25.67 is about at the 80th percentile for
  high school graduates - reading levels of
  individuals in organizations may be better or
  worse, depending upon the organization and job
• Assess learner satisfaction with different
  narration types - might affect marketability

41
Interested in Instructional Design and E-learning?
Become a Bronco and take courses with a really
cool dude!
42
And these, too..
43
Questions?
alyce.dickinson_at_wmich.edu wendy_at_behtech.com