Transcriptionless discourse analysis realising the clinicians dream

1

• Transcription-less discourse analysis
  realising the clinicians dream?

M Brady,1 L Armstrong, 1,2 C Mackenzie, 3 J
Norrie 4
1 Nursing Midwifery and Allied Health Professions
Research Unit 2 Speech and Language Therapy, NHS
Tayside 3 Division of Speech and Language
Therapy, University of Strathclyde 4 Centre for
Healthcare Randomised Trials (CHaRT) Health
Services Research Unit, Aberdeen University
Views expressed are those of the authors and not
necessarily those of the Chief Scientist Office
2
Background

• Discourse Analysis (DA) is essentially seen as
  clinical research tool rather than a clinical
  tool
• Transcription preceding DA seen as time consuming
• Recent work (turn-taking and gesture use) showed
  potential of transcription-less approach
• Overview of results with full details available
  in a forthcoming Aphasiology paper

3
Objective

• To demonstrate the reliability and validity of
  transcription-less DA

4
Participants

• 10 individuals living in Tayside (Scotland)
• Right handed
• Mild-moderate post-stroke aphasia
• Current or recently past speech and language
  therapy clients
• Audio- and video-recorded discourse samples

5
Discourse samples
6
Measures
7
... more measures
8
still more measures
9
Analyses

• Transcription-based
• Two experienced DA raters
• Used transcripts, recordings and forms
• Transcription-less
• Four final year SLT student raters (plus one
  withdrawal)
• Five hours training
• Used recordings and forms only
• Inter-rater and intra-rater reliability
• Criterion validity (trans-based v trans-less
  analyses)

10
Results

• Transcription based analysis
• Intra-rater reliability no significant
  differences
• Inter-rater reliability no significant
  differences
• Transcription-less analyses
• Validity measured by comparing two approaches
• Reliability
• Continuous measures - intra-class correlation
  coefficient
• Categorical measures - percentage agreement.

11
Validity

• No significant differences between the scores
  from the two methods.

12
Reliability

• Inter-rater reliability
• Acceptable in general
• Strongest for gesture totals
• Varied for turn-taking and repair measures
• Categorical measures
• Initiation (topic and conversational)
• Concept Analysis
• Percentage agreement was good

13
Results Gesture Use (totals)
14
Results Topic Use (totals)
15
Results Turn-taking (totals)
16
Results Repair
17
Conclusions

• Demonstrated potential of transcription-less
  approach
• Valid Approach
• None of the measures overall gave significant
  differences between scores
• Armstrong, L, Brady, M, Mackenzie, C,
  Norrie, J (in press) Transcription-less analysis
  of aphasic discourse a clinicians dream or a
  possibility? Aphasiology.

18
Acknowledgements

• Participants
• Forth Valley Speakability
• New and Emerging Applications of Technology,
  Department of Health, UK
• Transcription-less raters
• Claire Higgins, Dorothy Russell, Kirsty
  McLaughlan, Lesley Garret and Sharon Nelson
• Transcription work - Caitriona Hutton

19

• Transcription-less discourse analysis
  realising the clinicians dream?

M Brady,1 L Armstrong, 1,2 C Mackenzie, 3 J
Norrie 4
1 Nursing Midwifery and Allied Health Professions
Research Unit 2 Speech and Language Therapy, NHS
Tayside 3 Division of Speech and Language
Therapy, University of Strathclyde 4 Centre for
Healthcare Randomised Trials (CHaRT) Health
Services Research Unit, Aberdeen University
Views expressed are those of the authors and not
necessarily those of the Chief Scientist Office
20
still more measures
21
Results Gesture Use (totals)
22
Topic use validity and reliability measures for
the transcription-less method