Abstract
The Attitudes toward Cooperative Industrial Robots Questionnaire (ACIR-Q) is a psychometrically validated, multidimensional self-report instrument designed by Leichtmann, Hartung, Wilhelm, and Nitsch (2023) to assess shop-floor manufacturing workers’ attitudes toward collaborative robots (commonly referred to as “cobots”). Developed in response to the rapid automation and digital transformation characterizing Industry 4.0, the ACIR-Q captures the multidimensional nature of human-robot collaboration across three core subscales: Affect-Behavior, Social beliefs, and Task beliefs. Drawing from foundational organizational psychology item pools by Chao and Kozlowski (1986) and Herold, Farmer, and Mobley (1985), the scale was refined and optimized using Ant Colony Optimization (ACO)—a computational meta-heuristic algorithm that isolates the most informative, psychometrically robust items while minimizing respondent burden. The final 12-item instrument utilizes a 5-point Likert scale (1 = Strongly disagree to 5 = Strongly agree) and was rigorously validated on German manufacturing personnel across both online and on-site industrial field samples. Confirmatory factor analyses demonstrated superior model fit statistics (Comparative Fit Index [CFI] ≈ 1.00; Root Mean Square Error of Approximation [RMSEA] ≈ 0.00), coupled with high internal consistency across all subscales (McDonald’s ω ranging from .81 to .84). The ACIR-Q offers researchers, industrial-organizational psychologists, and operations managers an empirical diagnostic tool to anticipate workforce resistance, inform sociotechnical system design, guide change-management initiatives, and monitor employee psychological wellbeing during the deployment of mobile and stationary cooperative robotics.
Keywords
Attitudes toward cooperative industrial robots, ACIR-Q, cobots, human-robot collaboration, sociotechnical systems, ant colony optimization, occupational psychology, technology acceptance, Industry 4.0, psychometrics
Authors
The Attitudes toward Cooperative Industrial Robots Questionnaire (ACIR-Q) was conceptualized, developed, and empirically validated by an interdisciplinary team of researchers in robopsychology, psychological assessment, and industrial engineering:
- Benedikt Leichtmann, Ph.D. (Corresponding Author)
Affiliation: Johannes Kepler University Linz, Department of Robopsychology, Linz, Austria.
ORCID: 0000-0002-9062-0996
Contact: [email protected] - Johanna Hartung, Ph.D.
Affiliation: University of Bonn, Department of Psychology, Bonn, Germany.
ORCID: 0000-0002-6392-4468
Contact: [email protected] - Oliver Wilhelm, Ph.D.
Affiliation: Ulm University, Institute of Psychology and Education, Ulm, Germany.
ORCID: 0000-0001-7980-1166
Contact: [email protected] - Verena Nitsch, Ph.D.
Affiliation: RWTH Aachen University, Institute of Industrial Engineering and Ergonomics, Aachen, Germany.
Contact: [email protected]
Purpose
The deployment of cooperative industrial robots (cobots) represents a fundamental paradigm shift in modern manufacturing. Unlike traditional industrial robots that operate enclosed within safety cages due to severe kinetic risks, collaborative robots share unsegregated physical workspaces with human workers. This intimate physical proximity, coupled with direct dynamic coordination on assembly lines, introduces complex psychological stressors, safety anxieties, and social disruptions within factory teams. Historically, many industrial robotics initiatives have suffered from low employee adoption, sabotage, heightened occupational strain, and increased turnover intention because organizational interventions ignored shop-floor attitudes.
The primary purpose of the ACIR-Q is to serve as a precise, brief diagnostic instrument to evaluate manufacturing workers’ multidimensional attitudes toward the integration of cooperative robots. Rather than treating “technology acceptance” as an abstract, unitary construct, the ACIR-Q dissects employee perceptions into affective-behavioral inclinations, broader social-relational dynamics, and task-specific functional utilities. This granular diagnosis enables industrial-organizational psychologists and operations managers to pinpoint exact sources of apprehension or enthusiasm within specific production units.
In applied organizational settings, the ACIR-Q provides actionable baseline data prior to the procurement and programming of robotic solutions. For example, if a department exhibits exceptionally low scores on the Task subscale, managerial interventions can focus on practical ergonomics, mechanical reliability, and technical training. Conversely, if low scores emerge on the Social subscale, management can address existential threats regarding job displacement, social isolation, and disruption of peer communication. In academic and empirical research, the ACIR-Q provides a standardized, psychometrically validated benchmark for testing sociotechnical theories, ergonomics models, and human-robot interaction (HRI) dynamics within real-world factory ecosystems.
Psychological Construct
Attitudes toward emerging workplace technologies are fundamentally complex constructs that synthesize cognitive appraisals, affective responses, and behavioral intentions. Grounded in contemporary psychometrics, the ACIR-Q operationalizes attitudes toward collaborative robotics across three distinct yet interrelated psychological dimensions:
1. Affect-Behavior Dimension
The Affect-Behavior dimension captures workers’ visceral, emotional reactions and personal behavioral readiness regarding close-contact collaboration with an industrial robot. Unlike abstract cognitive appraisals of utility, this dimension gauges immediate psychosomatic states such as nervousness, apprehension, unease, and comfort. Psychological research demonstrates that working in unseparated physical proximity with heavy, autonomous mechatronic arms frequently triggers heightened baseline vigilance and fear of mechanical malfunction or physical collision. In the ACIR-Q, this construct reflects both the negative valence of robotic anxiety and the positive valence of anticipation, behavioral comfort, and willingness to adapt personal workflows alongside the machine.
2. Social Beliefs Dimension
The Social Beliefs dimension examines the broader systemic, interpersonal, and sociotechnical ramifications of automated agents entering human work groups. Manufacturing floors are rich social environments where peer communication, informal mutual assistance, shared professional identity, and perceived organizational value provide essential psychological resources. Cobots can inadvertently act as social barriers—physically partitioning workers, modifying ambient acoustic profiles, reducing spontaneous interpersonal interaction, and eliciting fears of replacement or technological obsolescence. This dimension assesses perceived erosion of job meaningfulness, existential threats regarding long-term job security, loss of worker self-worth, and disruption of collaborative team cohesion.
3. Task Beliefs Dimension
The Task Beliefs dimension evaluates functional, instrumental, and ergonomic utility appraisals. Rooted in cognitive evaluations of workplace efficiency, this dimension assesses whether employees perceive the cobot as a genuine functional asset or as an operational hindrance. Specific facets include the clarity of reorganized workflows, occupational safety improvements, reduction of physical strain (e.g., musculoskeletal relief via ergonomic load balancing), and the robot’s capacity to streamline complex or physically demanding tasks. High scores on this dimension signify that employees view the collaborative robot not as a competitor or an opaque obstacle, but as a supportive tool that enhances day-to-day productivity and physical wellbeing.
Theoretical Framework
The ACIR-Q is structurally anchored in several foundational paradigms across psychological, ergonomic, and organizational literature:
The Tripartite Model of Attitudes
The core conceptual architecture of the ACIR-Q originates from the classical tripartite model of attitudes (Rosenberg & Hovland, 1960; Eagly & Chaiken, 1993), which posits that attitudes comprise affective (feeling states), cognitive (beliefs and evaluations), and behavioral (intentions and action tendencies) components. While many traditional technology assessment scales conflate these categories into an undifferentiated “favorable vs. unfavorable” continuum, Leichtmann et al. (2023) demonstrated that in human-robot collaboration, affective-behavioral inclinations function in tandem with distinct cognitive-evaluative frameworks (divided into task-specific functional beliefs and broader social-environmental beliefs).
Sociotechnical Systems Theory (STS)
Originally formulated by Eric Trist and Ken Bamforth (1951) at the Tavistock Institute, Sociotechnical Systems Theory asserts that organizational performance depends on the joint optimization of both the technical subsystem (machinery, automation, algorithms) and the social subsystem (people, relationships, authority structures, social values). When advanced collaborative robots are introduced strictly as technical optimizations without accounting for the social fabric of the shop floor, psychological friction and organizational resistance inevitably emerge. The ACIR-Q’s inclusion of a dedicated Social dimension directly reflects STS principles by measuring the perceived impact of automation on human social capital, workplace meaning, and collective peer dynamics.
Technology Acceptance Model (TAM) and Job Demands-Resources (JD-R)
While Fred Davis’s (1989) Technology Acceptance Model highlights Perceived Usefulness (PU) and Perceived Ease of Use (PEOU)—constructs mirrored within the ACIR-Q’s Task dimension—the scale also integrates insights from the Job Demands-Resources (JD-R) model (Demerouti et al., 2001). Under the JD-R framework, cobots can function simultaneously as a resource (alleviating physically demanding labor, enhancing safety) and as a demand (generating cognitive load, introducing anxiety, disrupting interpersonal support). The ACIR-Q precisely quantifies these opposing psychological forces.
Validity
The construct, test, and criterion validity of the ACIR-Q were comprehensively demonstrated across multiple phases of scale construction and empirical testing by Leichtmann et al. (2023):
Construct Validity
Construct validity was established through sequential exploratory factor analysis (EFA) and confirmatory factor analysis (CFA) across diverse cohorts of German industrial manufacturing employees, including crowdsourced online samples and in-situ field samples from manufacturing plants. Across confirmatory evaluations, the hypothesized three-factor structure (Affect-Behavior, Social beliefs, Task beliefs) demonstrated exceptional fit indices that significantly outperformed alternative one-factor and two-factor models.
Convergent and Discriminant Validity
Convergent validity was demonstrated through substantial factor loadings across all 12 items on their respective latent dimensions, with standardized factor loadings consistently exceeding .60, and the vast majority falling between .70 and .85. The three latent factors exhibited moderate inter-factor correlations (typically ranging from r = .45 to .65), confirming that while the subscales share a common higher-order orientation toward automated technology, each captures substantial, non-redundant variance. Discriminant validity was supported by demonstrating that task beliefs remain empirically separable from emotional anxieties and broader social disruption concerns.
Criterion-Related and Predictive Validity
The ACIR-Q demonstrates robust criterion-related validity in relation to real-world behavioral and psychological outcomes. Total and subscale scores correlate significantly with workers’ willingness to participate in cobot pilot training, subjective physiological strain during robot collaboration trials, generalized technology commitment, and trust in workplace automation. Furthermore, negative scores on the Social beliefs subscale reliably predict turnover intentions and organizational cynicism following technology announcements.
Reliability
The ACIR-Q demonstrates exceptional psychometric reliability across independent manufacturing cohorts. Rather than relying solely on Cronbach’s alpha (α), which is known to be biased under conditions of varying factor loadings, the authors computed McDonald’s omega (ω), representing a more accurate assessment of internal consistency within multidimensional frameworks:
- Overall Scale Reliability: Across testing cohorts, the composite scale demonstrated composite reliability coefficients (ω) ranging from .81 to .84, indicating high internal stability.
- Affect-Behavior Subscale: Demonstrated strong internal consistency, with McDonald’s ω values consistently observed around .82.
- Social Beliefs Subscale: Displayed robust internal coherence, with ω values reaching approximately .84.
- Task Beliefs Subscale: Maintained high internal consistency, with ω values of approximately .81.
Test reliability was further corroborated through rigorous item-total correlation assessments, where every selected item demonstrated corrected item-total correlations exceeding .45. The stability of the parameter estimates across both online validation samples and physical factory field deployments demonstrates that the instrument is free from substantial measurement error and displays stable performance across varied industrial settings.
Factor Analysis
The structural composition of the ACIR-Q was established through an innovative psychometric procedure combining Exploratory Factor Analysis (EFA), Confirmatory Factor Analysis (CFA), and algorithmic Ant Colony Optimization (ACO):
Initial Item Pool and Exploratory Factor Analysis (EFA)
The scale development began with an extensive initial item pool derived from validated scales by Chao and Kozlowski (1986) and Herold, Farmer, and Mobley (1985). An initial EFA on the candidate items identified five distinct factors explaining approximately 42% of the cumulative variance. However, a pressure-related factor accounted for only 3% of the total variance and exhibited weak, unstable factor loadings. Consequently, this dimension was pruned to enhance structural clarity, yielding a theoretical three-factor architecture encompassing affective/behavioral inclinations, social/organizational beliefs, and task/ergonomic beliefs.
Ant Colony Optimization (ACO) Item Selection
To eliminate redundant items and prevent questionnaire fatigue among shop-floor personnel, Leichtmann et al. implemented Ant Colony Optimization (ACO). ACO is an advanced meta-heuristic algorithm that simulates the foraging behavior of ant colonies to solve combinatorial optimization problems. In psychometrics, ACO algorithms systematically explore vast configurations of item subsets to identify the precise combination of items that maximizes measurement quality—optimizing model fit, maximizing latent factor variance, and ensuring balanced subscale lengths—while maintaining comprehensive construct coverage. The ACO algorithm converged on an optimal 12-item solution (4 items per factor).
Confirmatory Factor Analysis (CFA) Fit Statistics
The resulting 12-item, three-factor structure was evaluated using CFA in an independent field validation sample. The model demonstrated extraordinary goodness-of-fit metrics:
- Comparative Fit Index (CFI): ≈ 1.00
- Tucker-Lewis Index (TLI): ≈ 1.00
- Root Mean Square Error of Approximation (RMSEA): ≈ 0.00
- Standardized Root Mean Square Residual (SRMR): ≤ .03
These empirical metrics indicate an exceptionally close correspondence between the theoretical three-factor measurement model and the observed covariance structure of the industrial sample.
Instrument / Measurement Tool
- Full Instrument Name: Attitudes toward Cooperative Industrial Robots Questionnaire (ACIR-Q)
- Authors: Benedikt Leichtmann, Johanna Hartung, Oliver Wilhelm, and Verena Nitsch (2023)
- Assessment Type: Self-report psychometric questionnaire
- Administration Format: Paper-and-pencil or computer-based / mobile survey
- Target Population: Manufacturing employees, assembly line technicians, and industrial operators (aged 18 years and older)
- Item Count: 12 items
- Subscales (4 items each):
- Affect-Behavior Subscale: Items 1, 2, 3, 4
- Social Beliefs Subscale: Items 5, 6, 7, 8
- Task Beliefs Subscale: Items 9, 10, 11, 12
- Authentic Response Scale: 5-point Likert scale:
- 1 = Strongly disagree
- 2 = Disagree
- 3 = Neither agree nor disagree
- 4 = Agree
- 5 = Strongly agree
- Scoring and Inversion Guidelines:
- Reverse-Coded Items: Items 1, 2, 5, 6, 7, and 8 are negatively phrased and must be reversed prior to scoring (i.e., 1 → 5, 2 → 4, 3 → 3, 4 → 2, 5 → 1).
- Subscale Scores: Calculated by computing the mean or sum of the 4 designated items per dimension after recoding reverse-worded items.
- Composite Score: An overall attitude index can be computed as the arithmetic mean of all 12 items. Higher scores consistently reflect more favorable, supportive, and constructive attitudes toward cooperative industrial robots.
- Completion Time: Approximately 3 to 5 minutes
Permissions & Fee and Test Year
The Attitudes toward Cooperative Industrial Robots Questionnaire (ACIR-Q) was officially published in 2023. In accordance with open science practices, the instrument is licensed under the Creative Commons Attribution 4.0 International (CC BY 4.0) license. The scale is freely accessible and commercially non-proprietary; no royalty fees, licensing costs, or formal developer approvals are required for educational, academic, or applied organizational research uses, provided appropriate scientific citation is granted to Leichtmann et al. (2023). Additional methodological appendices, item translations, and supplementary psychometric materials are accessible via the Open Science Framework (OSF) repository.
References
- Chao, G. T., & Kozlowski, S. W. (1986). Employee perceptions on the implementation of robotic manufacturing technology. Journal of Applied Psychology, 71(4), 652–659. https://doi.org/10.1037/0021-9010.71.4.652
- Davis, F. D. (1989). Perceived usefulness, perceived ease of use, and user acceptance of information technology. MIS Quarterly, 13(3), 319–340. https://doi.org/10.2307/249008
- Demerouti, E., Bakker, A. B., Nachreiner, F., & Schaufeli, W. B. (2001). The job demands-resources model of burnout. Journal of Applied Psychology, 86(3), 499–512. https://doi.org/10.1037/0021-9010.86.3.499
- Eagly, A. H., & Chaiken, S. (1993). The psychology of attitudes. Harcourt Brace Jovanovich College Publishers.
- Herold, D. M., Farmer, S. M., & Mobley, M. I. (1985). The human side of robotics: Results from a field study of employee reactions to the introduction of industrial robots. Georgia Institute of Technology.
- Leichtmann, B., Hartung, J., Wilhelm, O., & Nitsch, V. (2023). New short scale to measure workers’ attitudes toward the implementation of cooperative robots in industrial work settings: Instrument development and exploration of attitude structure. International Journal of Social Robotics, 15(6), 909–930. https://doi.org/10.1007/s12369-023-00996-0
- Rosenberg, M. J., & Hovland, C. I. (1960). Cognitive, affective, and behavioral components of attitudes. In M. J. Rosenberg & C. I. Hovland (Eds.), Attitude organization and change: An analysis of consistency among attitude components (pp. 1–14). Yale University Press.
- Trist, E. L., & Bamforth, K. W. (1951). Some social and psychological consequences of the longwall method of coal-getting: An examination of the psychological situation and defences of a work group in relation to the social structure and technological content of the work system. Human Relations, 4(1), 3–38. https://doi.org/10.1177/001872675100400101