Abstract
Students with attention-deficit-hyperactivity disorder (ADHD) have a variety of social and pragmatic language deficits. They have difficulty understanding contextual cues reflected by problems processing prosody, facial cues, and body gestures to interpret intentions and points of view of others. They exhibit problems initiating interactions, maintaining conversation, waiting their turn, and appropriately using tone and volume when speaking. They tend to be overly literal in their comprehension and have trouble understanding humor. Successful interactions require understanding of non-literal language, particularly by later elementary school. Verbal irony and sarcasm are used to indirectly convey attitudes or beliefs or to generate humor. In most situations in which one is sincere, there is a concordance between linguistics and affective prosody (e.g., sad words are paired with sad prosody), but in cases of sarcasm, the words and prosody differ. Research has shown that children with ADHD tend to have problems processing affective prosody. They are less accurate than typically developing children in matching affective prosody to affective facial expression (Shapiro et al., 1993).
Sarcasm is a type of speech expressed through contextual and prosodic cues. It is often characterized by a paralinguistic profile that signals to the listener to avoid the interpretation of the utterance as sincere. Because understanding of sarcasm relies on interpreting social cues, students with ADHD may have difficulty distinguishing between sarcasm and sincerity. The purpose of this study was to assess the ability of children with ADHD to recognize sincerity and sarcasm using the Awareness of Social Inference Test (TASIT; McDonald et al., 2002), a video tool designed to assess social perception with vignettes reflecting everyday expressions. The test consists of three kinds of exchanges: sincere, where the speaker means what he or she says; sarcastic, where the same ambiguous script is enacted in such a way that it is clear the speaker is meaning something quite different to that literally asserted; and paradoxical sarcasm, where the script literally makes no sense unless it is assumed one speaker is being sarcastic (McDonald, 2012).
The following three questions were addressed:
Do children diagnosed with ADHD show problems in their understanding of sarcasm?
If so, are the differences in their understanding of sarcasm dependent on the speakers’ actions, their speech, thoughts, and/or feelings?
Can individual symptoms of ADHD predict the ability to identify sarcasm?
Method and Material
Participants
A total of 22 children between ages 11.3 and 15 years with a clinical diagnosis of ADHD took part in the study. A control group was selected based on (a) the age of the children with ADHD and (b) no known diagnosis of any clinical disorder. The resulting groups were matched for gender, chronological age, and receptive language ability.
Assessments
The British Picture Vocabulary Scale–Third Edition (BPVS-III; Dunn et al., 2009) was used to assess receptive language ability.
The children completed Conners’ (2008) self-report short rating scales, a standardized assessment of ADHD symptoms appropriate for children and adolescents between 8 and 18 years of age. The short version was used, which consists of 39 questions in 5 key areas of interest: inattention, hyperactivity and impulsivity, learning problems, aggression, and family relations. Each question consists of a statement such as “I blurt out the first thing that I think of” or “I struggle to complete hard tasks.” Participants respond on a 4-point Likert-type scale where 1 indicates the person completely disagrees with the statement and a 4 indicates the person completely agrees with the statement.
Parents completed Conners’ (2008) Parent-Rating Full-length Scale for the ADHD children to confirm that all children reached the clinical level of symptoms required for a diagnosis of ADHD. The parents’ report provides an evaluation of the areas of inattention (likely to be inattentive, organizational problems, and problems finishing tasks); hyperactivity and impulsivity (inability to sit still, feel restless, and impulsive behavior); learning problems (struggling with school-based tasks); executive functioning (difficulties in planning, prioritizing, and organizing); aggression (difficulty in regulating their own emotions and prone to become irritable and angry easily); and peer relations (difficulty making and sustaining friendships). The parent circles a number on a 4-point Likert-type scale ranging from 0 to 3 for each statement, with a 0 representing “in the past month, this was not true at all about my child/It never happened” and a 3 representing “in the past month, this was very much true about my child/It happened very often.”
Sarcasm recognition was measured using Part 2 of the TASIT, which is composed of videotaped vignettes of everyday social interactions (McDonald et al., 2002). This test assesses participants’ understanding of conversational meanings that are determined by paralinguistic cues. It is composed of 15 short (15–60 s) vignettes that contain dialogues between two trained actors, conveying either a sincere or sarcastic exchange. Each of the vignettes was scored as 1 (correct) or 0 (incorrect) for identifying what the actor was doing, saying, thinking, and feeling, producing a maximum possible score of 60 (a maximum total score of 4 for each vignette). The scores can also be subdivided into three subscales corresponding to each type of exchange (sincere, simple sarcasm, and paradoxical sarcasm).
These videos test the student’s ability to perceive information that is implied rather than has been directly stated. Persons must distinguish sarcastic inference based on the appearance and overall manner of the actors, such as their facial expression, gestures, or tone of voice. For the sincere and simple sarcasm conditions, the content of the verbal script could be similar in content. For example, “I’d be happy to do it. I’ve got plenty of time” could be used as a script for each type. However, in the sincere verbal exchanges, the speakers mean what they are saying, that is, the words spoken and the paralinguistic cues are consistent, so the person is saying what the person means. Using the example script above, the correct interpretation for the sincere condition would be that the person was happy to help and felt they had the time. In contrast, for the simple sarcasm exchanges, one of the actors is being sarcastic (the literal meaning is different to the actual message the speaker is trying to express), but this can only be determined by reading the paralinguistic cues, such as facial expression, prosody of the voice, and posture of the body and hand, to identify the overall meaning of the speaker. Using the same script as before, for the simple sarcasm condition, the speaker may convey the same verbal content but would use exaggerated facial, body, and vocal language to show they meant the opposite (Rankin et al., 2009). Finally, in the paradoxical sarcasm exchanges, the dialogue between the two actors does not make sense unless sarcasm is recognized.
Procedure
All participants completed the BPVS first, followed by Conners’ self-rating scales, and finally, the TASIT. Before completing the TASIT, participants were told that they would watch a series of videos and they would have to answer a series of statements about each one of them. The statements of each video included descriptions of the actor’s specific intentions (“do” probes), what the actor wanted to verbally convey (“say” probes), what the actor was thinking (“think” probes), and the emotional state of actor (“feel” probes). An example of a question a participant would receive based on what the speaker wanted to convey is, “Is she trying to say she wanted to go out that night?” or based of what the speaker felt, “Is she happy to cancel?” Each statement was answered through a YES/NO response card. The answers determined whether the viewer was able to understand the meaning and intentions of the different exchanges.
Results
There were significant differences between the control group and the ADHD group on the self-rating scales of inattention, hyperactivity/impulsivity, learning problems, and aggression.
Performance across the three types of sarcasm and their different probes: The ADHD group exhibited significantly poorer performance in their overall identification of the sarcasm. For all students, the sincere exchanges were recognized significantly more often than simple sarcasm and paradoxical sarcasm; and simple sarcasm was identified more than paradoxical sarcasm. For all students, more exchanges were identified with the use of “do” probes compared with the “say” probes; and more exchanges were identified with the use of the “feel” probes compared with the “say” probes. There were no significant differences between the groups in the sincere exchanges or the simple sarcasm exchanges. However, there was a significant difference for the paradoxical sarcasm type, with the children with ADHD showing poorer performance. There were no significant differences between the groups on the “do,” “think,” or “feel” probes. However, there was a significant difference in the “say” probes, with the ADHD group identifying significantly fewer exchanges when asked to identify them based on what the person was saying.
The relationship between the symptoms of ADHD and performance across the TASIT: Multiple regressions were also carried out to address the relationship between ADHD symptomology (inattention, hyperactivity and impulsivity, learning problems, aggression, and family relations) and performance across the three subscales (sincere, simple, and paradoxical) of the TASIT. For the children with ADHD, none of the five symptoms were found to be predictors of sincere or paradoxical sarcasm. However, for simple sarcasm the overall model was significant. This analysis also revealed a negative significant effect of aggression.
There was no relationship between parents’ ratings of ADHD symptoms as predictors of sincere, simple sarcasm or paradoxical sarcasm.
For both the ADHD and the typically developing children, there was no relationship between students’ BPVS scores and chronological age as predictors of each of the three types of sarcasm.
Discussion
This study investigated the ability of children with ADHD to identify complex cues of social interactions in videos. Children with ADHD were found to be less accurate at identifying paradoxical sarcasm compared with the typically developing children. However, when no irony was intended and the verbal content matched the non-verbal content (sincere exchanges), or when it was clear that the speaker meant something different to what was being asserted (simple sarcasm), the children with ADHD performed at similar levels to those of the typically developing children.
The finding that the ADHD group was significantly poorer at identifying paradoxical sarcasm is congruent with previous research showing deficits of complex social and emotional cues (e.g., Da Fonseca et al., 2009; Yuill & Lyon, 2007). Although pragmatic language difficulties have commonly been reported in children with ADHD (Adams et al., 2005; Martin & McDonald, 2003), and have been shown to be as extreme as those found in children diagnosed with autism spectrum disorder (ASD; Bishop & Baird, 2001), little research has addressed their ability to comprehend complex social interchanges, such as sarcasm. When deficits in emotion understanding have often been considered, they have typically being tested using a simple emotion face recognition paradigm (Da Fonseca et al., 2009; Yuill & Lyon, 2007).
The performance of students with ADHD on the TASIT, which resembles everyday social situations, highlights the difficulties they have in recognizing social and contextual cues and shows that both verbal and non-verbal affects are important to be able to interpret sarcasm, particularly when there is a mismatch between what they see and what is being said. Only when the speaker used exaggerated facial, vocal, and body language to indicate sarcasm, as was the case in the simple sarcasm condition, were the children with ADHD able to perform similarly to the typically developing children. The type of probe question also differentiated the ADHD and the typically developing children. Although both groups were able to correctly infer sincerity and sarcasm based on the actions and thoughts of the speaker in each of the vignettes, the children with ADHD were found to be significantly poorer only when identifying paradoxical sarcasm based on the discourse and emotional state of the speaker. Being able to understand the speakers’ intentions and beliefs are all characteristics of mentalizing abilities, and the results are therefore suggestive that the children with ADHD were able to make some inferences based on the speakers’ beliefs but struggled when making inferences based specifically on speakers’ feelings. In addition, the children with ADHD also showed poorer performance than the typically developing children in their understanding of what the person was saying, and this may indicate that problems in pragmatic language ability underlie some of their ability to distinguish between sincerity and sarcasm.
Students with ADHD may have social and pragmatic deficits as severe as those with ASD. The reasons for the children’s social difficulties may be misunderstood—they might be interpreted as due to failure to attend or to inhibit inappropriate behavior rather than as pragmatic language and theory of mind deficits. Children with ADHD are likely to have a meta-knowledge of social pragmatic rules, but in actual contexts they cannot read the social contextual cues and organize appropriate responses.