Kjelgaard & Tager-Flusberg (2013)


Prosodic Impairment Research Groups



ASD = autism spectrum disorders

C = Clinician

CA = chronological age

EBP = evidence-based practice

NA = not applicable

P = Patient or Participant

pmh = Patricia Hargrove, blog developer

SD = standard deviation

SLP = speech–language pathologist

SLI = speech-language impairment


SOURCE: Kjelgaard, M. M., & Tager-Flusberg, H. (2013). The perception of the relationship between affective prosody and the emotional content in utterances in children with autism spectrum disorders. Perspectives on Language Learning and Education, 20, 20-32. doi:10.1044/lle20.1.20




DATE: October 14, 2014

ASSIGNED GRADE FOR OVERALL QUALITY: B- (Highest possible grade is

B+ due to the design of the investigation.)


POPULATION: Autism spectrum disorders (ASD); Child


PURPOSE: To investigate the ability of children with ASD to perceive happy and sad emotions using prosody



  • Although children with ASD (mean IQ about 84, mean age 9-6) can attend to emotional prosody in comprehension task, they do experience difficulty.
  • The Ps with ASD have trouble when there is a mismatch between prosody and semantic content.
  • When there is a mismatch between prosody and words, Ps with ASD are more likely to attend to the words.
  • Overall, the investigators indicated that Ps with ASD had difficultly identifying (in a comprehension task) the emotional state of the speaker using prosody alone.
  • The investigators state that “there is a disconnect between the implicit processing of emotional prosody and the explicit labeling of the emotion in prosody” (p. 20).


  1. What type of evidence was identified? Prospective, Nonrandomized Group Comparison Design
  1. Group membership determination:
  2. If there were groups of participants, were members of groups matched? Yes
  3. The matching strategy involved
  • Participants (Ps) in the specific language impairment (SLI) and ASD groups were not significantly different with respect to chronological age (CA), overall IQ, verbal IQ, nonverbal IQ.
  • There were 2 other groups: young adults (YA) and typically developing (TD) children. Neither of these 2 groups were matched to each other or the SLI and ASD groups.
  1. Was participants’ communication status concealed?
  2. from participants? No
  3. from assessment administrators? No
  4. from data analyzers? No



  1. Were the groups/participants adequately described? Yes
  2. How many participants were involved in the study?
  • total # of participants: 102
  • was group membership maintained throughout the experiment? Yes
  • # of groups: 4
  • List names of groups:

     – ASD = children diagnosed with autism spectrum disorders

– SLT = children diagnosed with specific language impairment (>1 SD on formal test of language competence)

– TD = typically developing children

– YA = young adults associated with the lab and their colleagues

  • # of participants in each group:

     – ASD = 23

– SLI = 21

– TD = 35

– YA = 23


  1. The following variables were described
  • age:

     – ASD = mean age 9-6

– SLT = mean age 9-6

– TD = mean age 8-6

– YA = not provided

  • cognitive skills:

     – ASD = mean overall IQ 84.3, mean nonverbal IQ 87.7, mean verbal IQ 83

– SLT = mean overall IQ 97.2, mean nonverbal IQ 92.8, mean verbal IQ 88.5

– TD = not provided

– YA = not provided

  • expressive language: SLI group was <1 SD below the mean on a formal test of language


  1. Were the communication problems adequately described? No
  • disorder type: SLI, ASD; Child



  1. What were the different conditions for this research?
  2. Subject (Classification) Groups? Yes
  • ASD
  • SLI
  • TD
  • YA
  1. Experimental Conditions? Yes
  • affective prosody (happy, sad prosody)
  • semantic content (happy, sad sentence content)
  1. Criterion/Descriptive Conditions? No


  1. Were the groups controlled acceptably? Yes



  1. Were dependent measures appropriate and meaningful? Yes
  2. List dependent measures:
  • Dependent Measure #1: Response time to a 2-picture forced choice comprehension task  
  • Dependent Measure #2: Percentage of errors to a 2-picture forced choice comprehension task  
  1. Neither of the dependent measures are subjective.


  1. Both the dependent/ outcome measures are objective.



  1. Were reliability measures provided?
  2. Interobserver for analyzers? No
  3. Intraobserver for analyzers? No
  4. Treatment fidelity for investigators? No



  1. Description of design of study:
  • Investigators presented prerecorded audio recordings to Ps who were expected to point to the picture (of 2 pictures) that best represented the sentence.
  • The stimuli had been constructed to represented sentence with

– matching emotional prosodic and semantic content or

– mismatching emotional prosodic and semantic content.

  • The 3 emotional states were: happy, sad, and neutral.
  • The investigators insured that the semantic and prosodic content reflected the targets emotions using judges and for the prosodic content they also included acoustic analyses.
  • When the mismatch occurred Ps were instructed to attend to either the emotional or the prosodic content.
  • Due to matching strategies the investigators only undertook selected comparisons of the P groups: YA versus TD and SLI versus ASD.


 a.  What were the results of the inferential statistical testing?

The following comparisons were significant ( p ≤ 0.05):

  • Dependent Measure #1: Response time to a 2-picture forced choice comprehension task  

   – YA versus TD:

  • group main effect: TD significantly slower than YA
  • prosody x semantic interaction: both YA and TD were significantly slower for mismatches between emotional semantic and emotional prosodic content compared to matches between semantic and prosodic representations of emotions.

– SLI versus ASD:

  • prosody x semantic interaction: both SLI and ASD were significantly slower for mismatches between emotional semantic and emotional prosodic content compared to matches between semantic and prosodic representations of emotions.


  • Dependent Measure #2: Percentage of errors to a 2-picture forced choice comprehension task  

   – YA versus TD:

  • for the ignoring the voice (prosodic content) task group main effect: Overall both groups did very well judging the emotional semantic content.
  • for the ignoring the words (semantic content) task, there were several significant findings:
  1. YA were more accurate than TD
  2. TD had more trouble in mismatched semantic and prosodic contexts,
  3. TD had more trouble with the neutral conditions.

– SLI versus ASD:

  • for ignoring the voice/prosody task–prosody x semantic interaction and prosody x semantic x diagnosis interaction: Ps with ASD did better in matching (emotional semantic and prosodic) contexts than in mismatching contexts but Ps with SLI did not exhibit this pattern.
  • for ignoring the words/semantics task: there were several significant findings
  1. Ps with ASD had trouble ignoring the prosody. That is, they interpreted the words of the utterance to identify emotional content even when they were directed to attend to the prosody only.
  2. Ps with ASD they did better in matching than in neutral or mismatching contexts.
  3. Ps with ASD did not have similar problems.
  4. Overall, Ps with SLI did significantly better (i.e., fewer errors) than Ps with ASD.

b.  The statistical test used to determine significance was ANOVA

c.   Were effect sizes provided? No

d.  Were confidence interval (CI) provided? No





Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

%d bloggers like this: