 |
Choice Responding as a Function of Choice-Presentation Method and Level of
Preference in Persons with Developmental Disabilities
by Lisa Schwartzman, Dickie C. T. Yu, and Garry L. Martin
University of Manitoba and St. Amant Centre
Abstract
This study examined whether the discrimination skills of an individual, as measured by the Assessment of Basic Learning Abilities (ABLA) test, can be used to predict the optimal stimulus modality (objects, pictures, and spoken cues) for presenting choices. In addition, we examined whether the predictions would hold for both highly and moderately preferred items. Choices were presented to six adults with developmental disabilities using a two-choice format, and involving a high preference pair of food items (one item was chosen over the other approximately 70% of the time) and a low-preference pair of food items (one item was chosen over the other approximately 30% of the time). Each pair of items was presented to each participant in three modalities (objects, pictures, and spoken-only presentation) using a reversal design with replications. The participants' abilities to make consistent choices with the high preference pair was accurately predicted 100% of the time by their prior performance on the Assessment of Basic Learning Abilities (ABLA) test, and choice consistency was accurately predicted 88% of the time by the ABLA test with the low preference pair of food items. The results also showed that different preference levels, at least for the two levels studied, did influence the stability of choice responding. The results suggest that the ABLA test may be a useful tool for matching the choice presentation method to the discrimination skills of individuals with developmental disabilities.
Introduction
Preference assessment is often used to identify reinforcers for individuals with severe developmental disabilities (e.g., Green, Reid, Canipe, & Gardner, 1991; Green et. al., 1988; Parsons & Reid, 1990; Wacker, Berg, Wiggins, Muldoon, & Cavanaugh, 1985; Windsor, Piche, & Locke, 1994). For example, to determine whether a client prefers an apple or an orange, that client may be asked to choose between them when the actual items are placed in front of the client. But preference can be difficult to assess for activities involving less manipulable objects (e.g., watching a television set versus walking on a treadmill), and between events and activities that involve many stimuli and behaviors (e.g., deciding whether to go to a Mexican restaurant versus a Chinese restaurant). One option is to present choices via pictures. Another option is to describe choices verbally. Unfortunately, many persons with developmental disabilities are unable to demonstrate consistent preferences when options are presented as pictures or spoken descriptions (Parsons, Harper, Jensen, & Reid, 1997). Is it possible for staff working with individuals with developmental disabilities to accurately predict which choice presentation method (objects versus pictures versus spoken) that would be most appropriate for individual clients? And is the consistency of choice-making by different choice presentation methods affected by the degree of preference that a client has for one alternative over another? The present study examined these questions.
Conyers et al. (in press) used performance on the Assessment of Basic Learning Abilities (ABLA) test to accurately predict whether or not participants would be able to make consistent choices when options were presented as objects, pictures, or spoken descriptions. The ABLA test assesses the ease or difficulty with which a participant is able to learn a simple imitation (level 1), a two-choice visual discrimination where position is a relevant cue (level 2), a two-choice visual discrimination (level 3), a two-choice visual quasi-identity match-to-sample discrimination (level 4), a two-choice auditory discrimination with position as a relevant cue (level 5), and a two-choice auditory-visual combined discrimination (level 6) (Kerr, Meyerson, & Flora, 1977; Martin & Yu, 2000). The ABLA tasks have been shown to be hierarchically related, have high inter-tester and test-retest reliability, have high predictive validity, and can be administered in approximately 30 minutes (Martin & Yu, 2000). Conyers et al. showed that individuals who were able to perform a simple visual discrimination (ABLA level 3), but not matching (ABLA level 4) or auditory-visual discriminations (ABLA level 6), demonstrated consistent preferences with the actual items (objects), but not when the items were presented pictorially or as spoken descriptions; individuals who were able to perform ABLA level 3 and 4 discriminations demonstrated consistent preferences with objects and pictures of those objects, but not when the items were presented as spoken words; and individuals who were able to perform ABLA level 6 auditory-visual discriminations demonstrated consistent preferences with all three presentation methods. Therefore, failure to demonstrate a preference may be a result of not matching the stimulus presentation method to the client's discrimination ability, rather than an absence of preference between the available alternatives.
Although Conyers et al. (in press) replicated their results in two experiments, involving food and non-food items, respectively, the results were more consistent with food items. It was not clear, however, whether this was due to differences in reinforcing value (i.e., food items were more reinforcing than non-food items) or differences in some other characteristics of the two stimulus classes. In this study, we sought to systematically replicate the study by Conyers et al. to determine if the ABLA test could reliably predict how different choice presentation methods influence the choice responding of individuals with different discrimination skills. Moreover, we evaluated how choice responding is affected by more or less preferred stimuli in the same stimulus class using food items only.
Method
Participants and Setting
Participants were six adults (3 females and 3 males) living in a residential and community resource facility for persons with developmental disabilities. Written informed consent was obtained from the legal guardian of each participant prior to the study. The participants were selected based on their visual and auditory discrimination skills as measured by the ABLA test. Lea and Tom were 28 and 23 years old, respectively, and both were diagnosed as functioning at the severe level. Both passed the ABLA level 3, the visual discrimination task, but failed ABLA level 4 and level 6, the visual match-to-sample and auditory-visual discriminations. Mic and Bob were 28 and 44 years old, respectively, and both were diagnosed as functioning at the moderate level. Both passed ABLA levels 3 and 4, but failed ABLA level 6. Lastly, Thea and Laurie were 44 and 35 years old, respectively. Thea was diagnosed as functioning at the mild level and Laurie at the severe level. Both passed all three discriminations (ABLA levels 3, 4, and 6). Sessions were conducted in an assessment room. During all sessions, a participant sat across a table from the tester. An observer was present in the room during reliability assessments.
General Research Plan and Experimental Design
The research plan involved first screening potential participants using the ABLA test to determine their abilities to perform simple visual, match-to-sample and auditory-visual discriminations. Next, a preference assessment was conducted with each participant to identify food items that could be used in the high and low preference conditions. Lastly, the effects of three choice presentation methods and two preference conditions on choice responding were evaluated using a combined reversal and multi-element design (Martin & Pear, 1999). The reversal design was used to evaluate presentation methods (object, picture, and spoken) and each participant was exposed to the following phases: OPOSOPOSO (where O = object, P = picture, and S = spoken presentation phase). During each phase (presentation method), two preference conditions (high and low) were alternated in a multi-element design.
Discrimination Skills Assessment
Each participant was assessed using the ABLA test (Kerr et al., 1977). For the present study, the three discriminations of interest involved ABLA levels 3, 4, and 6. For the level 3 task, the participant was required to place a manipulandum (a piece of foam with neutral color) into the same container (e.g., yellow can) when presented with a yellow can and a red box in randomly alternated left-right positions. For the level 4 discrimination, the participant was required to put a manipulandum into the matching container (e.g., a yellow cylinder in the yellow can and a red cube in the red box). For the level 6 discrimination, the participant was required to place a piece of foam into the container that was verbally requested by the tester (e.g., "yellow can" or "red box"). As described by Martin and Yu (2000), correct responses were reinforced with praise and an edible and incorrect responses were followed by a correction procedure (a demonstration, guided practice, and an opportunity to respond independently). Eight consecutive correct responses defined a pass on a discrimination task and eight cumulative errors defined a fail.
Preference Assessment
A preference assessment among 12 food items was conducted using the paired-comparison procedure (Fisher et al., 1992). The food items were selected based on recommendations from caregivers, and availability and ease of presentation. An attempt was made to include both preferred and less preferred food items. The assessment consisted of 132 trials, conducted over several sessions, with each item being paired with every other item two times. Throughout the study, a choice response was defined as an individual touching or pointing to one of the two alternatives presented. On each trial, the participant was given the selected item to consume.
High and Low Preference Conditions
Following the preference assessment, items were ranked by the percentage of trials on which they were selected, and two pairs of
items were chosen for each participant. The most and least frequently chosen items formed the "high preference" pair (high preference
condition). The difference in preference between the two items in the high preference condition averaged 74% across the participants,
ranging from 59% to 83%. Two other items that were closer in preference were selected as the "low preference" pair (low preference
condition) for each participant. The difference in preference between the two items in the low preference condition averaged 33%
across the participants, ranging from 32% to 36%.
Presentation Methods
Object presentation
During the object presentation phase, the participant was prompted to sample each
food item at the beginning of each session. On each subsequent trial, the tester placed the two food items in front of and at equal
distance to the participant and asked him/her to "pick one", without saying the items' names. The participant received the selected
edible for consumption, and the item that was not chosen was removed from the table. After the edible had been consumed, a new trial
was presented. The right-left positions of the two alternatives were counterbalanced across trials. Within each 20-trial session, the
high and low preference conditions were alternated after every two trials, resulting in 10 trials per condition.
Picture presentation
During this phase, the sessions were conducted in the same manner as the object phase, except that realistic color pictures (approximately 20 cm by 25 cm) of the food items were presented instead of the actual items. At the beginning of each session, the participant was shown each picture and was then presented with the corresponding food item to consume. During each subsequent trial, the tester placed the two pictures in front of the participant and asked him/her to "pick one", without saying the items' names. The participant received the food item that corresponded to the selected picture, and the picture that was not selected was removed from the table. High and low preference conditions were alternated as in the object phase.
Spoken presentation
During this phase, the sessions were conducted in the same manner as the object and picture conditions, except that the tester presented the participant with two identical opaque containers, each of which concealed a food item. At the beginning of each session, the tester stated the name of an edible while pointing to the corresponding container concealing that edible, and then opened the container and gave the food item to the participant. This was then repeated with the second container. Thereafter, on each trial, the tester stated the name of each edible while pointing to the corresponding container, before asking the participant to "pick one". The order in which the names of the edibles were stated was counterbalanced across trials. The participant received the food item from the container that was selected on each trial, and the container not chosen was removed from the table. The high and low preference conditions were alternated as in the previous phases.
Inter-observer Reliability and Procedural Integrity
Inter-observer reliability and procedural integrity checks were conducted for each participant throughout the study. The percentage of sessions with reliability checks ranged from 38% to 83% across participants. During reliability checks, an observer independently recorded the participant's choice on each trial. A trial was considered an agreement if the tester and the observer recorded the same response, and a disagreement if they did not. Percent agreement was calculated for each session by dividing the number of agreements by the total number of agreements plus disagreements and multiplying by 100%. Agreement was 100% in all cases.
The observer also assessed procedural integrity using a checklist. The procedural checklist included: presenting the edibles in the correct positions, giving the appropriate instruction, allowing the participant 10 s to respond, repeating the verbal prompt if no response was made after 10 s, providing the participant the selected food item, and removing the food item that was not selected after a choice was made. A trial was considered to have been delivered correctly if all of the above steps occurred; otherwise, it was scored as incorrect. The percentage of correctly delivered trials was calculated for each session and procedural integrity was 100% in all observed sessions.
Results
Figure 1 shows the percentage of responses to the preferred item during high and low preference conditions across presentation methods. During high preference conditions, all participants displayed clear preferences (70% or higher) consistent with their prior preference assessments only when the presentation modality required a discrimination skill that they had demonstrated on the ABLA test. For example, Lea and Tom, who passed ABLA level 3 (visual discrimination), but not ABLA level 4 (visual match-to-sample) or ABLA level 6 (auditory-visual discrimination), showed clear preferences consistent with their preference assessments during the object presentation phases, but not during picture or spoken presentations. Mic and Bob, who passed both ABLA levels 3 and 4, but not level 6, showed clear preferences consistent with prior assessments during both object and picture presentations, but not during spoken presentations. Lastly, Thea and Laurie, who passed all three discriminations on the ABLA test, showed clear preferences consistent with prior assessments during all three methods of presentation.
Figure 1. Percent of responses to preferred stimuli during high and low preference conditions across object (O), picture (P), and spoken (S) choice presentation methods.
Three patterns of responding could be seen during the low preference condition. First, for Lea and Bob, the response patterns during low preference trials were similar to the high preference condition (i.e., consistent with prior preference assessment) across all phases, although their preferences were weaker in the low preference condition than in the high preference condition during some phases (second object phase for Lea and in most object and picture phases for Bob). Second, Mic, Thea, and Laurie showed the same patterns in both the high and low preference conditions in only some of the phases. During the low preference condition, Mic's preference was near chance level during the second picture phase, Thea's preference was near chance level during the first object phase, and Laurie's preference was near chance level during both the spoken phases. The third pattern of responding, exhibited by Tom, showed no clear preference in the low preference condition across all phases.
Figure 2 shows the mean percentage of responses to the preferred stimuli during the high and low preference conditions and for each presentation method. In the high preference condition, the ABLA test performance accurately predicted responding in all 18 instances (6 participants x 3 presentation methods), with Lea and Tom (the level 3 participants) showing a consistent preference in the object phase only, Mic and Bob (the level 4 participants) showing a consistent preferences in the object and picture phases only, and Thea and Laurie (the level 6 participants) showing consistent preference for object, picture, and spoken phases. In the low preference condition, the ABLA test accurately predicted performance in 16 of 18 instances, although consistency of preference was frequently less than in the high preference condition. In two instances in the low preference condition (Tom in the object presentation and Laurie in spoken presentation), a preference was not demonstrated.
Figure 2. Mean percent of responses to preferred stimuli during high and low preference conditions during each presentation method.
Discussion
The results add to previous research (Conyers et al., in press) that the ABLA test is a more accurate predictor of discrimination ability than is IQ. For Lea and Tom, who passed only up to ABLA level 3, their ABLA performance accurately predicted that they would show consistent choices only in the object phase. For Laurie, who passed ABLA level 6, her ABLA level performance accurately predicted consistent choices in the object, picture, and spoken phases. Yet all three participants were diagnosed as functioning in the severe level of mental retardation.
An interesting response pattern observed during this study for both individuals who were capable of discriminating between objects and pictures (e.g., Mic and Bob who passed ABLA level 4) and for one of the two individuals capable of discriminating between objects, pictures, and spoken cues (Laurie who passed ABLA level 6), was that their preferences, when demonstrated, were usually slightly higher or more stable for objects than for pictures or spoken presentation within each preference condition. Since object presentation requires a simple simultaneous visual discrimination while picture and spoken presentations require a conditional discrimination, it is possible that preferences are more likely to be affected when the discriminations required to demonstrate those preferences are procedurally more complex. In addition, more consistent preferences might be expected from object presentation since the food objects are unconditioned reinforcers, whereas the pictures and spoken cues are conditioned reinforcers (paired with food).
The present study defined high and low preference conditions by the relative distance between the two choice stimuli, expressed as percent selection during the preference assessment. Future study is needed to examine how the reinforcing value of a stimulus (expressed as percent preference) may be a function of the reinforcing value of other available options. That is, the reinforcing value of popcorn, for example, may be very high when presented along with carrots, but much lower when presented along with potato chips.
A limitation of the current and the Conyers et al. study (in press) is that both studies used within-subjects designs, involving a small number of participants. Research with additional participants and with group designs will extend the generality of these findings. Another limitation of the present study is that the assessment stimuli were confined to food reinforcers. Although Conyers et al. examined food and non-food reinforcers, the latter were confined to simple activities. Since many decisions that impact on quality of life are abstract and involve protracted leisure activities, future research is needed to determine whether basic discrimination skills, such as those measured by the ABLA test, are reliable predictors of stimulus modality for choice presentations involving complex activities.
Nevertheless, the present findings underscore the need to select the stimulus modality that is matched to the individual's discrimination skills for presenting choices. As noted by Conyers et al. (in press), when the presentation method requires a discrimination skill not yet demonstrated by the individual (e.g., presenting choices with spoken cues to persons who have demonstrated only visual discriminations on the ABLA test), a caregiver may erroneously conclude that the person has no preference or is unable to make choices. The former may lead to the selection of an ineffective reinforcer in training programs and the latter may limit future choice opportunities for the person. Given that research has shown that staff can readily learn to administer the ABLA test through self-instruction and that the test can be administered in a brief period of time (DeWiele, Martin, & Garinger, 2000), caregivers should consider assessing the discrimination repertoire of individuals with disabilities to determine the optimal stimulus modality for preference assessment and choice presentations. Another advantage of assessing the discrimination repertoire of individuals is that discriminations that have not yet been learned (e.g., auditory discrimination) could be targeted for training. Moving from object to picture to spoken modalities will enhance the independence of the person as well as the practicality of choice presentation for the caregiver.
References
Conyers, C.M., Doole, A., Vause, T., Harapiak, S., Yu, D., & Martin, G.L. (in press). Predicting the optimal choice-presentation method for assessing choices of persons with developmental disabilities, Journal of Applied Behavior Analysis.
DeWiele, L., Martin, G.L., & Grainger, J. (2000). Field testing of a self-instructional manual for the ABLA test. Journal on Developmental Disabilities, 7, 93-108.
Fisher, W. W., Piazza, C. C., Bowman, L. G., Hagopian, L. P., Owens, J. C., & Slevin, I. (1992). A comparison of two approaches for identifying reinforcers for persons with severe and profound disabilities. Journal of Applied Behavior Analysis, 25, 491-498.
Green, C. W., Reid, D. H., Canipe, V. S., & Gardner, S. M. (1991). A comprehensive evaluation of reinforcer identification processes for persons with profound multiple handicaps. Journal of Applied Behavior Analysis, 24, 537-552.
Green, C. W., Reid, D. H., White, L. K., Halford, R. C., Brittain, D. P., & Gardner, S. M. (1988). Identifying the reinforcers for persons with profound handicaps: Staff opinions versus systematic assessment of preferences. Journal of Applied Behavior Analysis, 21, 31-43.
Hughes, C., Hwang, B., Kim, J.H., Eisenman, L.T., & Killian, D.J. (1995). Quality of life in applied research: A review and analysis of empirical measures. American Journal on Mental Retardation, 99, 623-641.
Kerr, N., Meyerson, L., & Flora, J. (1977). The measurement of motor, visual and auditory discrimination skills. Rehabilitation Psychology, 24, 95-112.
Martin, G.L., & Pear, J.J. (1999). Behavior modification: What it is and how to do it, 6th Ed. Upper Saddle River, NJ: Prentice-Hall.
Martin, G.L., & Yu, D. (2000). Overview of research on the Assessment of Basic Learning Abilities test [Special Issue]. Journal on Developmental Disabilities, 7(2), 10-36.
Parsons, M. B., Harper, V. N., Jensen, J. M., & Reid, D. H. (1997). Assisting older adults with severe disabilities in expressing leisure preferences: A protocol for determining choice-making skills. Research in Developmental Disabilities, 18, 113-126.
Parsons, M. B., & Reid, D. H. (1990). Assessing food preference among persons with profound mental retardation: Providing opportunities to make choices. Journal of Applied Behavior Analysis, 23, 183-195.
Wacker, D. P., Berg, W. K., Wiggins, B., Muldoon, M., & Cavanaugh, J. (1985). Evaluation of reinforcer preferences for profoundly handicapped students. Journal of Applied Behavior Analysis, 18, 173-178.
Windsor, J., Piche, L.M., Locke, P.A. (1994). Preference testing: A comparison of two presentation methods. Research in Developmental Disabilities, 15, 439-455.
Acknowledgements
We thank the participants and staff for their cooperation throughout the study. Brenda Van Wyck, Ryan Goodman and Shayla Harapiak assisted on reliability checks. This research was supported by Medical Research Council of Canada grant #MOP-36433, and has received ethical approval from the Education Nursing Research Ethics Board, University of Manitoba.
Submitted by
Lisa Schwartzman, Dickie C. T. Yu and Garry L. Martin
Authors' Biographies
Lisa Schwartzman is a graduate student in the Department of Psychology at
the University of Manitoba, specializing in applied behavior analysis.
Email: lisasc@mts.net
Dickie Yu is the Director of Research at the St. Amant Centre, and an Adjunct
Professor in the Faculty of Nursing at the University of Manitoba.
He has published widely in the field of developmental disabilities.
Email: yu@stamant.mb.ca
Garry Martin is a Professor of Psychology at the University of Manitoba. He has
co-authored numerous books and many journal articles in behavioral
psychology, many of which focused on research in developmental disabilities.
Email: gmartin@cc.UManitoba.CA
Mailing Addresses
D. Yu, St. Amant Centre, 440 River Road, Winnipeg, MB, Canada R2M 3Z9
G. Martin, Department of Psychology, University of Manitoba, 129 St. Paul's College, Winnipeg, MB, Canada, R3T 2M6
|
