Rule-based insensitivity to contingencies according to self-reported generalized pliance/tracking and the instruction functions

Context and Objectives

Rule-governed behavior is a fundamental phenomenon in the analysis of human behavior and constitutes an active area of research within Relational Frame Theory (RFT). One of the most significant findings in this research tradition is rule-based insensitivity to contingencies, that is, the tendency of individuals to persist in responding according to previously learned rules, even when environmental contingencies have changed and these responses are no longer functional. This phenomenon has been documented in multiple experimental studies and has important implications both for the theoretical understanding of human behavior and for clinical and educational applications.

Within research on rule-governed behavior, Relational Frame Theory has identified two primary functions of instructions: pliance and tracking. Pliance refers to the behavior of following instructions motivated by the social or contextual consequences associated with compliance with the rule, independent of whether those consequences derive from a direct correspondence with reality. Tracking, in contrast, describes the following of instructions motivated by the correspondence between the instructional response and the natural contingencies of the environment. These two forms of instructional control may have distinct implications for sensitivity to changes in environmental contingencies.

The article by Martínez-Carrillo and Ruiz (2026), published in the Journal of Contextual Behavioral Science, addresses a novel research question: Do individuals who predominantly present generalized pliance differ from those who show generalized tracking in their sensitivity to changes in contingencies when under the control of explicit instructions? Specifically, the authors aimed to analyze how the interaction between instruction type (general instructions versus general instructions plus explicit response criteria) and the predominant style of rule-following (pliance or tracking) affects insensitivity to contingencies.

The general objective of the study was to examine, through a 2×2 factorial design, how sensitivity to contingencies varies as a function of: (a) the type of instructional function (predominant pliance versus tracking) that participants self-report in standardized questionnaires, and (b) the type of instruction manipulated experimentally (general instructions versus general instructions plus explicit response criteria). This research represents an important advance as it is, according to the authors, the first experimental study that selects participants specifically based on their predominant self-reported type of rule-following.

Method

Participants and Selection Models

The study included 90 undergraduate students selected from an initial population of 565 participants who completed screening questionnaires. Participant age ranged from 18 to 50 years, with a mean of 20.44 years (SD = 3.27). The demographic composition was predominantly female: 58 participants (64.44%) were women, 31 (34.44%) men, and 1 (1.11%) identified as non-binary.

Participants were assigned to two groups based on their scores on two standardized self-report instruments: the Generalized Pliance Questionnaire (GPQ) and the Generalized Tracking Questionnaire (GTQ). The High Pliance group (n = 45) showed high scores on the GPQ (M = 87.18, SD = 10.99) and low scores on the GTQ (M = 44.62, SD = 7.35). The High Tracking group (n = 45) showed the inverse pattern: low GPQ scores (M = 35.92, SD = 11.03) and high GTQ scores (M = 69.24, SD = 3.92). The GPQ consists of 18 items on a 7-point Likert scale with internal consistency of α = 0.94, while the GTQ contains 11 items with the same scale and α = 0.93.

Experimental Design

The study employed a 2×2 factorial design combining two independent variables: (1) predominant type of rule-following (High Pliance vs. High Tracking) and (2) type of instruction (General Instructions [GI] vs. General Instructions + Response Criteria [GI + RC]). This combination generated four experimental conditions: (a) High Pliance with GI, (b) High Pliance with GI + RC, (c) High Tracking with GI, and (d) High Tracking with GI + RC.

Materials and Experimental Task

The experimental task was a computerized matching-to-sample (MTS) task using quadgrams (meaningless consonant groupings such as GNHV, GNHN, GNKS, RWZF). The task was programmed in Visual Basic 6.0 on computers with 22-inch screens.

The task was divided into two blocks of 60 trials each, totaling 120 trials:

Block 1: Participants were instructed to select the quadgram most similar to the sample stimulus AND respond slowly (within the last 5 seconds of the available period).

Block 2: The contingencies were modified without prior notice. Participants now had to select the quadgram most different (opposite to the previous instruction) AND respond quickly (within the first 5 seconds of the available period).

In the GI condition, participants received general instructions about the task without explicitly specifying the correct response criteria. In the GI + RC condition, the same general instructions were provided BUT explicit rules were added that specified that they should respond SLOWLY and select the most SIMILAR option.

Procedure and Dependent Variables

After completing the task, participants responded to the Experimental Task Survey (ETS), a structured instrument that assessed functions related to instructions and the point system through specific questions.

The dependent variables were:

Contingency sensitivity: Measured as the number of trials required to emit 3 consecutive correct responses that differed from the response pattern established in Block 1.
Rule persistence: Measured as the number of trials in Block 2 during which participants continued to respond according to the Block 1 pattern (i.e., selecting similar and responding slowly).

Data Analysis

Data analysis was performed using JASP version 0.19.3.0. Given the characteristics of the data, non-parametric tests were employed, including likelihood ratio tests, Kruskal-Wallis tests for between-group comparisons, and Mann-Whitney tests for pairwise comparisons. The study was approved by the Institutional Ethics Committee of Fundación Universitaria Konrad Lorenz.

Compensation and Context

Participants received compensation of 10,000 Colombian pesos (approximately 2.5 USD) for their participation. The study was conducted in a computer room with approximately 30 computers available.

Results

Block 1: Acquisition of Initial Response

In Block 1, 14 of the 90 participants (15.55%) did not meet the specified mastery criterion. There was a significant difference between instruction conditions: 10 participants in the GI condition and only 4 in the GI + RC condition failed to meet the criterion (χ²(1) = 4.83, p = .03). This suggests that the inclusion of explicit response criteria facilitated correct acquisition of the required response pattern in Block 1.

Regarding the number of trials required to meet the criterion, the GI + RC condition required significantly more trials (M = 55.48, SD = 12.07) compared to the GI condition (M = 49.81, SD = 11.09), representing a statistically significant difference (H(1) = 24.11, p < .001). However, there was no significant effect of predominant type of rule-following (H(1) = 1.08, p = .30), indicating that both High Pliance and High Tracking participants required similar numbers of trials.

It is important to note that 7 additional participants in the GI + RC condition did not respond correctly to the initial trials and were withdrawn from the analysis. This reduced the final sample for the main analyses to 71 participants (78.89% of the original sample).

Contingency Sensitivity

The analysis of contingency sensitivity (number of trials required to change response) showed a statistically significant effect of instruction type (H(1) = 13.88, p < .001). Specifically, participants in the GI + RC condition required significantly more trials to change their responses (M = 9.13, SD = 14.41) compared to those in the GI condition (M = 5.07, SD = 11.60). This result confirms the classic effect of rule-based insensitivity to contingencies: when explicit instructions with specified response criteria are provided, participants show greater difficulty in detecting and adapting to changes in environmental contingencies.

There was no significant effect of predominant type of rule-following (H(1) = 0.82, p = .37), indicating that both High Pliance and High Tracking participants showed similar patterns of insensitivity to contingencies when under the control of explicit instructions. This finding was relatively surprising, as the authors had predicted that High Pliance participants would show greater insensitivity to contingencies.

Rule Persistence

Rule persistence in Block 2 (number of trials during which participants continued to respond according to Block 1) showed a marginally significant effect of instruction type (H(1) = 3.66, p = .056). Participants in the GI + RC condition showed numerically greater persistence (M = 12.28, SD = 17.31) compared to those in the GI condition (M = 6.23, SD = 10.16), although this difference was only marginally significant.

Again, there was no significant effect of predominant type of rule-following (H(1) = 0.27, p = .60), suggesting that instruction type was the determining factor of rule persistence, independent of the predominant style of rule-following reported by participants.

Experimental Task Survey: Functions of Instructions

The analyses of the Experimental Task Survey did not reveal significant differences among the four experimental conditions regarding functions related to instructions or the point system. However, when the rule groups (High Pliance vs. High Tracking) were compared independent of condition, important differences emerged:

High Pliance participants reported significantly higher scores on instruction-related functions (H(1) = 3.98, p = .046) compared to High Tracking participants. Similarly, High Pliance participants reported higher scores on point-related functions (H(1) = 5.54, p = .02).

Reasons for Continuing to Respond According to Block 1

Significant differences were found between conditions in two specific questions from the Experimental Task Survey that assessed reasons for continuing to respond according to Block 1:

Question 2 ("The instructions told me to respond slowly and select the most similar"): H(1) = 9.24, p = .002. Participants in the GI + RC condition endorsed this reason more frequently, confirming that they indeed followed the explicit instructions provided.

Question 3 ("I wanted to complete the task as instructed"): H(1) = 12.52, p < .001. Again, participants in the GI + RC condition endorsed this reason more frequently.

An additional analysis conducted specifically in GI + RC participants revealed a particularly interesting finding: among those who responded affirmatively to the question "I wanted to complete the task as instructed," there was significantly LOWER contingency sensitivity (M = 5.93, SD = 3.25) compared to those who responded negatively (M = 3.00, SD = 1.41). Similarly, those who responded affirmatively showed GREATER rule persistence (M = 14.29, SD = 16.59) compared to those who responded negatively (M = 2.60, SD = 1.52). These differences were statistically significant (Mann-Whitney test: U = 14.50, p = .03 for sensitivity; U = 16.00, p = .04 for persistence).

Motivations Related to Social Appearance

Regarding motivations to not change responses in Block 2, High Pliance participants reported significantly greater motivation derived from "not wanting the researcher to know I changed" (H(1) = 5.60, p = .018) compared to High Tracking participants. This finding is consistent with the theoretical conceptualization of pliance as behavior motivated by social consequences.

Reasons for Changing in Block 2

The analyses revealed significant differences by type of rule-following in the reasons for changing responses in Block 2:

Question 1 ("I wanted to earn points"): H(1) = 3.90, p = .048.

Question 3 ("I didn't want to be the only one not earning points"): H(1) = 15.44, p < .001. This effect was particularly pronounced, suggesting that consideration of social norms or expectations played an important role in participants' decisions to change responses.

Question 4 ("I didn't want the researcher to know I wasn't earning points"): H(1) = 14.26, p < .001. This finding also reinforces the importance of social consequences in participants' behavior.

Additional Observations: False Memory of Instructions

An additional finding of clinical importance was that approximately 30-40% of participants in the GI condition (who did not receive explicit response criteria) incorrectly believed they had been instructed about what the correct responses should be. This false memory phenomenon was associated with lower contingency sensitivity and greater rule persistence, suggesting that even belief in instructions that were not actually provided can affect behavior.

Discussion and Conclusions

The results of this study replicate and extend previous findings on rule-based insensitivity to contingencies, confirming that when explicit instructions with specified response criteria are provided, individuals tend to persist in responding according to those rules, even when environmental contingencies have changed. This replication is consistent with the systematic review by Kissi et al. (2020) and with decades of research in behavior analysis and contextual behavioral science.

However, the most surprising finding of the study was that the insensitivity to contingencies effect was NOT significantly moderated by the predominant self-reported type of rule-following (pliance versus tracking). The authors had initially predicted that High Pliance participants would show greater insensitivity to contingencies compared to High Tracking participants, based on the assumption that pliance is more dependent on social context and less sensitive to the correspondence between response and natural contingencies. However, both groups showed similar patterns.

The authors propose a post-hoc explanation for this surprising result: it is possible that the greater motivation of High Pliance participants for instruction-related functions was compensated by their also greater motivation for functions related to point attainment. In this way, although High Pliance participants may have been more sensitive to the instructional aspects of the task, their greater motivation to earn points (an economic consequence) may have counterbalanced any advantage in terms of response flexibility.

The qualitative analyses of the reasons reported by participants provide additional insight. Between 70-80% of participants in the GI + RC condition reported that they continued responding according to Block 1 because "the instructions told me to respond slowly and select the most similar," directly confirming that these participants were governed by the explicit instructions. In contrast, only 30-40% of participants in the GI condition provided similar reasons, confirming that this condition did not produce the same level of explicit instructional control.

The study also documented the presence of false memory: approximately 30% of participants in the GI condition incorrectly believed they had been instructed about the correct response criteria, despite the fact that the GI condition specifically did not provide this information. Interestingly, these participants showed patterns of contingency sensitivity and rule persistence more similar to those of the GI + RC condition, suggesting that belief in instructions (even if false) can have functional effects similar to actual instructions.

The authors acknowledge several study limitations. First, the sample consisted solely of university students, limiting generalization to other populations. Second, the study was conducted in a Colombian population, which may limit the cross-cultural applicability of findings. Third, a dropout rate of 21.11% was observed, primarily due to participants who did not respond correctly to the initial trials in the GI + RC condition. Fourth, the authors themselves recognize that the general instructions included the phrase "the task may vary," which may have particularly sensitized High Pliance participants to the possibility of contingency changes.

Despite these limitations, the study makes significant contributions. It is the first experimental study to select participants specifically based on their predominant self-reported type of rule-following, providing a methodology for future research. Furthermore, the study demonstrates the importance of analyzing not only whether insensitivity to contingencies effects occur, but also understanding the motivational processes (instruction-related functions versus point-related functions) that underlie these effects.

Significance and contribution

This study contributes to understanding rule-governed behavior by demonstrating that the predominant type of rule-following (pliance versus tracking) has important implications for sensitivity to contingency changes under explicit instructions. The findings show that individuals with elevated pliance tendencies may be more vulnerable to rule-based contingency insensitivity, while those with elevated tracking tendencies maintain greater sensitivity to environmental changes. These results have implications for understanding behavioral rigidity and for designing interventions that promote greater behavioral flexibility in educational, clinical, and occupational contexts.