Journal of Undergraduate Research
Volume 5, Issue 1 - October 2003

Parent-Directed Behavioral Outliers: Their Cause and Result In Final Data Manipulation 

Paula Modesitt

ABSTRACT

This paper reviews the disorder of autism, the significance of parent-directed behavior to improve autistic children’s responses, coding procedures utilized by the data collection team, and the analysis of outliers identified in various parent-children sessions. The results include presentation of the total number of outlying sessions as well as close inspection of one family’s parent-child interactions. In addition to interesting findings, results of this work illustrate the value of single-subject experimental methodology in the studying complex behavioral disorders and parent-child social interactions.

BACKGROUND AND SIGNIFICANCE

Background and Significance


Diagnosis of Autism. Autism, previously regarded as a bewildering and debilitating disorder associated with mental retardation and impaired social abilities, has evolved into a more understood disorder where behavior can be positively altered towards the social norm with proper guidance. Characteristics of autism include impairments in social interaction, impairments in verbal and nonverbal communication, and restricted and repetitive patterns of behavior (American Psychiatric Association, 1994). The cause of autism, which emerges either at birth or early childhood, has not been identified. However there is evidence suggesting neurobiological and/or genetic components. In order to diagnose autism, deficits in social and communication as well as repetitive movements should be present by 36 months of age, according to the World Health Organization (WHO; 1993). Other characteristics vary with the age of the child, and many children with autism remain apparently content to be left alone to engage in self-stimulatory and other unusual activities. Older children fail to engage in usual forms of social interchange and have difficulties taking another person’s point of view into account (Volkmar and Cohen, 1985).


Prevalence. Autism occurs in one (1) in every five-hundred (500) births and in a rate of five (5) boys to every girl (Middlebrook, 1998, ¶1). Diagnosing this disorder can be difficult because at ages younger than two, an “atypical social interaction” as classified by the parents could really just be a child’s normal development into a socially shy, introverted personality.

Other disorders are closely linked with autism, as 75% of children diagnosed with autism are also mentally retarded, with IQ scores less than 70. However, many people have heard of the child with autism who possesses an incredible talent with music, math, or the arts. The range of these skills tends to be quite narrow and usually involves memory for concrete events of trivial information (e.g. numbers, calendars), visual-spatial or manipulative skills (e.g. jigsaw puzzles), or a precocious interest in letters and numbers (Volkmar and Cohen, 1985).

METHODOLOGY

Defining and Coding Behaviors

Before conducting any experiment involving behavior, it is important that the investigator understands exactly what behaviors are under investigation. As an Undergraduate Research Scholar, I participated in a study of autistic children where fathers were instructed by Jennifer Elder, PhD, RN, to elicit important social and pre-language behaviors from their children. Using the father-directed teaching techniques, fathers were taught skills including “imitation with animation” (IA) and “expectant waiting” (EW). During IA, fathers directly copied, in an animated manner, their children’s behavior, including noises, gestures, facial expressions, and intelligible words. When performing expectant waiting, fathers asked the children question, or performed certain behaviors to elicit a response. Then the fathers waited 3-5 seconds before repeating this same behavior. This particular waiting skill is important because parents of autistic children often become frustrated when the children do not answer right away, or divert their attention to another activity. This creates more parental stress, and parents often respond by increasing their questioning. This is actually counter-productive in that it makes the communication process even worse, since the autistic children need more time to react to questions and commands. By waiting 3-5 seconds before cuing the children again, parents give the children time to grasp the concept being initiated, and can respond in a more socially acceptable manner.

After the fathers mastered IA and EW, they were instructed to teach their wives these two behaviors. Part of this experiment was not only to test the children’s reaction to the fathers’ behavior, but to see if the mothers could be taught to carry out these social interaction modifications with their children, and elicit a positive social change in their children as well.

These father and mother play sessions were videotaped, and coders were responsible for interpreting the videotapes by logging each response and reaction into a statistic-compiling computer program. On the computer, codes 1 - 6 corresponded with “Parent Initiations”, “Parent Responses”, “Child Initiations” “Child Responses” Parental Imitation, and Parental Expectant Waiting. Codes 7 and 8 corresponded to vocal utterances (unintelligible uttering) and stereotypy, which are repetitive, apparently nonfunctional movements of body parts or manipulations of objects (e.g. body-rocking, head-rolling, spinning/shaking objects) (Bodfish, et al, 1995). Codes 9 and 0 corresponded to “Intelligible Words” and “Tantrums”, and Code T was used when a turn was completed. (A turn consists of an appropriate response following an initiation from either the child or the parent.)

After each entry, the computer program calculated the frequency of each code. These numbers were then compared from before the father-directed behaviors were modified, and before the father taught the mother these same behaviors. After analyzing data for 18 parent-child dyads, it was noted that the children’s initiations, responses, and intelligible words all significantly increased (p< .05), and there was no significant change in stereotypy frequencies from baseline to post-intervention conditions (p< .30).

Identification of Outlying Sessions

However, using visual analysis common to single subject experimental methodology, it was also noted that some of the general trends were interrupted by occasional outlying sessions. An outlier was an interruption in a stable trend, or a deviation from an otherwise upward or downward trend. Therefore, a random decrease or increase in any of the behaviors indicated an outlier. Outlying sessions were reviewed to determine what factors might have contributed to the aberrant frequencies. It is very important to determine what caused the outliers, because they have the possibility of skewing the overall results and comprising project internal validity

RESULTS

Upon close inspection, it became clear that certain behaviors were associated with more outliers than others. Seventy-five (75) of the total (approximately 700) sessions were identified as outliers. The behavioral outliers are described in Table 1 and presented in order of frequency. For example, child vocalizations were extremely variable and produced a total of ten (10) outlying sessions. Next were the father/mother responses, which totaled nine (9) outlying sessions. Child responses, IA, and EW all yielded eight (8) outlying sessions. Next came the father/mother initiations with a total of seven (7) outliers, and child initiations had a total of five (5) outlying sessions. The incidences of these outlying sessions, as well as the operational definitions of the behaviors, are displayed in Table 1. The number of outlying sessions dominated when the behavior “Expectant Waiting” was tested. This yielded seven (7) outlying father sessions, yet only one (1) outlying mother session. This difference between father and mother dyads could have some clinical significance, considering the fathers were the ones taught this behavior from the expert. Tantrum behaviors were seen in four (4) of the fathers’ outlying sessions, but in zero (0) of the mothers’.

Table 1
Operational Definitions of Behavioral Responses and Number of Outliers for Study Dyads (N=36)
Behaviors Operational Definition
Outliers (Father)
Outliers (Mother)
Father/mother initiating The movement cycle that begins with the parent orienting toward, reaching for, talking to, or offering items to the child and ends when the child has an opportunity to respond to the parent's initiation.
3
4
Father/mother responding The occurrence of a child behavior and ends after the parent initiates the requested movement or offers a comment that indicates that the parent acknowledges, understands, and/or approves of the child imitation or behavior within 2 seconds.
5
4
Child initiating The movement cycle that begins with a child orienting toward, reaching for, talking to, or offering items to the parent and ends when the parent has an opportunity to respond to the parent's initiation.
2
3
Child responding The movement cycle that begins with the occurrence of a parental prompt and ends after the child initiates the behavior within 5 seconds.
4
4
Imitation with Animation The movement cycle that begins within 5 seconds of a child's initiation wherein the parent imitates the child's behavior in an animated manner.
6
2
Expectant Waiting The movement cycle that begins when the parent prompts a child behavior and waits at least 5 seconds, during which time the parent provides the child with facial positions signaling the availability of positive social interactions and ends with the desired child response or a parental prompt if the response does not occur.
7
1
Child vocalizations Audible vocal sounds emitted by the child excluding laughing, crying, coughing, yawning, or sneezing and bounded by a parental vocal response or a 2-second pause.
5
5
Child stereotypies Repetitive, apparently nonfunctional, movements of body parts or manipulations of objects (e.g., body-rocking, head-rolling, spinning/shaking objects)
4
7
Child intelligible words Utterances that can be clearly understood by a rater who views the videotaped parent-child sessions and was previously unfamiliar with the child.
1
4
Father tantrums Clearly audible crying sounds emitted by the child, associated with kicking and/or flailing arms and bounded by a 5 second pause
4
0
Father/mother initiated turn completed The movement cycle that begins when the father initiates either a verbal or physical cue to the child and ends when the child responds to that cue within 5 seconds in a positive, expected manner.
0
2
Total Number of Outlying Sessions  
75


Outlying Sessions for Family T

Space limitations prohibit close examination of all individual family data. Therefore, I have chosen to highlight Family T’s to illustrate how outliers appeared and were determined because after visual inspection of all of the families’ graphs, it was noted that Family T had a large number of outlying sessions. Six (6) outlying sessions were identified for this family, and the tapes were reanalyzed. The certain outlying behaviors specific to this family are displayed in Table 2.

Table 2
Operational Definitions of Behavioral Responses of Outlying Sessions for Family T (N=1)
Behaviors Operational Definition
Outliers (Father)
Outliers (Mother)
Father/mother initiating The movement cycle that begins with the parent orienting toward, reaching for, talking to, or offering items to the child and ends when the child has an opportunity to respond to the parent's initiation.
1
0
Father/mother responding The occurrence of a child behavior and ends after the parent initiates the requested movement or offers a comment that indicates that the parent acknowledges, understands, and/or approves of the child imitation or behavior within 2 seconds.
1
0
Expectant Waiting The movement cycle that begins when the parent prompts a child behavior and waits at least 5 seconds, during which time the parent provides the child with facial positions signaling the availability of positive social interactions and ends with the desired child response or a parental prompt if the response does not occur.
1
0
Child vocalizations Audible vocal sounds emitted by the child excluding laughing, crying, coughing, yawning, or sneezing and bounded by a parental vocal response or a 2-second pause.
0
1
Child stereotypies Repetitive, apparently nonfunctional, movements of body parts or manipulations of objects (e.g., body-rocking, head-rolling, spinning/shaking objects)
0
1
Child intelligible words Utterances that can be clearly understood by a rater who views the videotaped parent-child sessions and was previously unfamiliar with the child.
0
1
Total Number of Outlying Sessions  
3
3


The three (3) outlying behaviors specific to the child-mother sessions in this family include: child vocalizations, which had a dramatic increase during the last session; stereotypy, which had a major deviation from the usual baseline of zero (0) tantrums; and intelligible words, in which there was a marked increase during the first intervention session. Figure 1 displays one of these graphs.

Figure 1. Example of one outlying session with mother
Figure 1. Example of one outlying session with mother.


The three (3) outlying behaviors specific to the child-father sessions in this family include: father initiation, which had one value with a marked deviation from the baseline; father responses, which had a dramatic dip in an otherwise slight downward trend; and father expectant waiting, which displayed one value as a dramatic increase from the almost stable baseline. Figure 2 displays one of these graphs.



Figure 2. Example of one outlying session with father.

All of the outlying sessions between the mother-child dyad yielded outlying sessions for child behavior. In contrast, all of the outlying sessions between the father-child dyad yielded outlying sessions for fathers’ behaviors. This could be attributed to many factors, but further research and examination is necessary pinpoint the exact cause. Plausible explanations include the child was having an “off” day during the outlying sessions, and the mother’s behavior in fact did not have any impact on the child’s responses. Or, perhaps the mothers’ behaviors towards their children were more stable than fathers’. This may be due to the fact that the mothers in each family spend more time with the autistic children and may have well-established and stable interaction patterns.

All of the outlying father-child sessions displayed outlying father behaviors. Perhaps when performing different interventions, fathers had to greatly modify their normal behavior with their children thus creating outliers. None of the fathers’ behaviors elicited an outlying response from the child, suggesting that the father-directed outliers are a result from the behavioral modification of the father.

DISCUSSION

Data Clinical Significance

The outlying sessions specific to the child-father interactions differed. These outlying sessions include: father initiating, which had a dramatic drop from the general trend line during the second intervention session; father responding, which had a drop in the first expectant waiting session; and expectant waiting, which had a jump in the third expectant waiting session.

After analyzing these specific sessions more closely, I determined that there were 3 (three) outlying sessions dealing with the child-mother interactions. However, during the child-father interactions, the only outlying sessions were outliers of the father’s behaviors. All of the outlying sessions with child-mother interactions deal with the child’s behavior as being skewed from the norm. However, all of the outlying sessions with child-father interactions deal seem to be more affected by deviations in the father’s behavior. This does not show that the fathers’ interventions created outliers in the children’s behavior, but that the interventions could have contributed to outlying father behaviors. It is not known whether the interventions caused the fathers to act differently in those sessions, or even if the interventions elicited outlying responses from the child in the child-mother sessions. A more clear inspection of each individual session as well as interpretation of the causes of the outlying sessions would be necessary to determine why these outlying sessions occurred.

Implications for Future Research

Due to the variability noted, it is critical to examine multiple sessions rather than engage in the common practice of solely comparing isolated pre and post-intervention data. It also appears that many factors may contribute to variability and adequate controls must be implemented. Finally, social reciprocity of father-child and mother-child interactions warrants further examination. That is, does variability in parent skill frequency directly affect child behaviors and does child behavior influence the frequency of parental skill use? Clearly more research is needed to examine these complex and important parent-child interactions in this fascinating population of children.

REFERENCES

Rawlings, American Psychiatric Association. (1994). Diagnostic and statistical manual of mental disorders (4th ed., revised). Washington, DC: Author.

Bodfish, J., Crawford, T., Powell, S., Parker, D., Golden, R., & Lewis, M. (1995). Compulsions in adults with mental retardation: prevalence, phemomenology, and comorbidity with stereotypy and self injury. American Journal of Mental Retardation, 175.

Middlebrook, B., Middlebrook, J., Adams, A. (1998,). Quick facts about Autism. Autism Info.com. Retrieved from http://www.autisminfo.com/QuickFacts.htm

Volkmar, F., Klin, A. & Cohen, D. (1997). Diagnosis and classification of autism and related conditiong: Consensus and issues. In D.J. Cohen & F.R. Volkmar (Eds.), Handbook of autism and pervasive developmental disorders (2nd ed; pp. 179, 419). New York: John Wiley & Sons.

World Health Organization. (1993). The ICD-10 classification of mental and behavioral disorders. Diagnostic criteria for research. Geneva, Switzerland: Author.

--top--

Back to the Journal of Undergraduate Research