Shapers@Work
Better living through Behavior Analysis
Verbal Operant Experimental (VOX) Analysis
General Procedures
A verbal operant experimental (VOX) analysis is a functional analysis of verbal behavior formalized for statistical analysis to answers the question: What controls the speaker's responding?
Research on verbal behavior has identified four primary sources of control over language:
Tact (i.e., labeling), controlled by the physical environment (Horne & Lowe, 1996)
Mand (i.e., requesting), controlled by restricted access (Michael, 1982a)
Intraverbal:
Echoic (i.e., echoing), controlled by an imitative verbal stimulus (Michael, 1982b)
Sequelic (i.e., conversing), controlled by other people's verbal behavior (Vargas, 1982)
In contrast to fluent speakers, who show proportionality across these four sources of control (Sundberg, 2007), speakers with stimulus overselectivity show different levels of response strength (Cochran, 1950; Lovaas et al., 1979). A VOX analysis identifies the extent to which the speaking repertoire is balanced.
Autistic Speaking Repertoire
Fluent Speaking Repertoire
The VOX analysis is composed of two to four assessment rounds, each consisting of three unique stimuli. The order of the assessment is as follows:
Round 1 (Items 1-3):
Tact
Mand
Echoic
Sequelic
Round 2 (Items 4-6)
Tact
Sequelic
Mand
Echoic
Round 3 (Items 7-9)
Tact
Echoic
Sequelic
Mand
Round 4 (Items 10-12)
Tact
Mand
Sequelic
Echoic
Power Analysis
After a sufficient sample size has been collected (i.e., 24 to 48 total responses), the VB-SCoRE can be calculated. The speaker record below can be used for data collection for each condition, decision making on how to proceed with the assessment, and data entry for the report generator.
Note: When marking the speaker record, "D" indicates that the referent served as a discriminative stimulus for the speaker's response, and "Δ" indicates no discriminative control over the speaker's verbal behavior.
Assessing the Ability to Label
The speaker’s ability to label is assessed as part of a free-operant preference assessment (Orsini et al, 2019). Throughout this condition, the speaker has the opportunity to choose from many different items with which to engage. Each round of the assessment consists of three unique items. Once an item has been selected, the speaker is asked to label the item (e.g., “What do you have there?”).
Apart from its use in assessing the speaker’s ability to label, the free-operant preference assessment ensures random sampling (i.e., the selection is determined by the speaker) and affords cultural sensitivity (e.g., the speaker, and/or their guardian, selects the assessment materials and provides the target response). The targets identified in the labeling condition are then also assessed under each of the remaining conditions.
Table 1 shows tact control assessed for three items assessed in Round 1 (items 1, 2, and 3), and three different items assessed in Round 2 (items 4, 5, and 6). Note that the three items in Round 1 would also be assessed for mand, echoic, and sequelic control prior to beginning the second round of the assessment. Between 6 and 12 total labeling responses are assessed in this condition.
Assessing the Ability to Request
The speaker’s ability to request is assessed as part of a multiple stimulus without replacement (DeLeon & Iwata, 1996) or paired-stimulus preference assessment (Fisher et al., 1992). Each of the three items identified in the labeling condition is systematically assessed as a request. Once an item has been selected, all other toys are removed. After the speaker has accessed the preferred object for 30 sec, that item is also removed to evoke a request from the child (e.g., “What do you want?”). This process is repeated until a request for all items has been assessed.
Table 2 shows mand control assessed for the same three items that were identified in each round of the tact condition. Note that each round of the assessment begins with assessing tact control. The order of the remaining conditions varies for each round of the assessment. Between 6 and 12 total requesting responses are assessed in this condition.
Assessing the Ability to Echo
The speaker’s ability to echo is assessed through discrete-trial training (Bogin et al., 2010), again using each of the three response targets from the previous conditions. At 30-sec intervals, the assessor requests for the speaker to copy what they say (e.g., Say, “Apple.”). This process is repeated until echoing the name of all items has been assessed.
Table 3 shows echoic control assessed for the same three items that were identified in each round of the tact condition. Note that each round of the assessment begins with assessing tact control. The order of the remaining conditions varies for each round of the assessment. Between 6 and 12 total echoing responses are assessed in this condition.
Assessing the Ability to Converse
Similar to the echoing condition, the speaker’s ability to converse is also assessed through discrete-trial training (Bogin et al., 2010) with each of the same three target responses. At 30-sec intervals, the assessor provides a fill-in-the-blank frame (e.g., “You eat an …”) or wh- question (e.g., "What do you eat?") for each of the targets. This process is repeated until an episode for each of the items has been assessed.
Table 4 shows sequelic control assessed for the same three items that were identified in each round of the tact condition. Note that each round of the assessment begins with assessing tact control. The order of the remaining conditions varies for each round of the assessment. Between 6 and 12 total conversing responses are assessed in this condition.