Personal Computing Utilization - Thompson et al. (1991)

Theory Publication Information

Citation Information

Why Was the Model Created?

Thompson, Higgins, and Howell developed their PC utilization model to address a significant gap in understanding personal computer adoption within organizational settings. While prior technology acceptance models existed, there was insufficient theoretical grounding specifically examining factors influencing actual PC usage patterns among end-users. Personal computing was becoming increasingly prevalent in organizational environments, yet organizations struggled to understand why some employees readily adopted PCs while others resisted or underutilized them.

Building on Triandis' (1971; 1980) theory of interpersonal behavior, the authors recognized that technology adoption exists on a continuum of utilization intensity rather than as binary use versus non-use. Prior research had focused heavily on intention to use or acceptance, but Thompson's work directly examined actual usage behavior and factors predicting variation in usage levels. The model proposed that six primary constructs predict PC utilization: social factors, affect, facilitating conditions, complexity, job fit, and long-term consequences.

The study surveyed knowledge workers across nine divisions of a large multinational manufacturing organization, yielding a final sample of 212 respondents (from 278 returned questionnaires). Findings were directly relevant to managers seeking to understand employee PC adoption patterns. By examining actual utilization rather than intentions, the model addressed the known intention-behavior gap that affected earlier theoretical frameworks. The authors explicitly recognized that technology adoption is multifaceted, requiring understanding of affective, social, cognitive, and contextual influences simultaneously.

Core Concepts and Definitions

The PC Utilization Model operationalizes seven primary constructs:

• Complexity: Perceived difficulty of using a PC. Measured through items assessing ease versus difficulty in learning and using computers.
• Job Fit: Perceived alignment between PC functionality and job requirements. Measured through items assessing whether PCs help with job performance and match job tasks.
• Long-Term Consequences: Perceived future payoffs from PC use including career advancement and productivity gains. Measured through items assessing beliefs about future benefits and career impacts.
• Affect: Emotional attitudes toward PCs. Measured through items assessing whether individuals like or dislike PCs and emotional responses.
• Social Factors:Perceived importance of others' opinions regarding PC use and social norms about technology. Measured through items assessing social influence and normative pressures.
• Facilitating Conditions: Objective and perceived availability of resources supporting PC use. Measured through items assessing training availability, equipment access, and technical support.
• Utilization: Actual frequency and intensity of PC usage. Measured through items assessing direct usage behavior including number of facility visits and frequency of use.

Preceding Models or Theories

The Thompson model built upon several prior intellectual traditions:

• Triandis' theory of interpersonal behavior (1971; 1980): Foundational framework positing that behavior is determined by intentions, habits, and facilitating conditions, with intentions shaped by social factors, affect, and perceived consequences.
• Technology Acceptance Model (Davis): Prior IS acceptance research providing context for investigating behavior prediction.
• Expectancy theory: Grounded the concept that expected consequences influence behavior.
• Porter and Lawler's theory of motivation: Provided foundations for understanding effort-performance-reward relationships.
• Theory of Reasoned Action (Fishbein & Ajzen, 1975): The dominant IS adoption framework at the time. Thompson et al. positioned their Triandis-based model as a competing alternative to TRA-based approaches.

Describe The Model

The PC Utilization Model specifies complex relationships where multiple constructs determine actual usage behavior. Expected consequences (complexity, job fit, long-term consequences) operate alongside social factors, affect, and facilitating conditions to predict utilization. The model explained 24% of variance in PC utilization through these multi-level constructs. Path analysis revealed that job fit and social factors showed strongest effects on utilization (.26 and .22 respectively).

What does the model measure?

• Expected consequences constructs: Complexity perceptions, job fit alignment, long-term consequence beliefs predicting utilization indirectly.
• Affective and social dimensions: Emotional attitudes and social normative pressures influencing utilization.
• Organizational support: Facilitating conditions reflecting technical support, training, and resource availability.
• Actual usage behavior: Direct measurement of utilization frequency, facility visit numbers, and time spent using systems.

Main Strengths

• Actual behavior measurement: Unlike many models measuring intentions, this model directly measures utilization behavior.
• Comprehensive construct coverage: Integrates affective, social, cognitive, and contextual factors providing holistic view of usage determinants.
• Grounding in established theory:Anchored in Triandis' theory providing theoretical justification beyond empirical discovery.
• Empirical strength: Explained substantial variance in utilization (R² = .24) demonstrating identified factors capture meaningful drivers.
• Clear managerial implications: Identifying that social factors and job fit are strongest predictors guides evidence-based intervention strategy.
• Integration of multiple levels: Synthesizes individual attitudes, social context, job characteristics, and organizational support systems.
• Sophisticated methodology: Used PLS analysis allowing simultaneous model testing with measurement error accounting.

Main Weaknesses

• Single-organization design: All data came from one organization, severely constraining generalizability to other organizations and contexts.
• Limited facilitating conditions operationalization: Measured narrowly as technical support, missing other resource factors like equipment access and software upgrade ease.
• Low reliability for complexity scale:Complexity had lowest Cronbach's alpha (.60), below conventional .70 threshold, indicating measurement issues.
• Non-significant relationships: Affect and facilitating conditions showed non-significant direct effects, contrasting with prior technology acceptance research.
• Self-report measurement: All data were self-reported rather than based on objective usage statistics, risking over- or underestimation of actual usage.
• Cross-sectional design: Single-point-in-time snapshot prevents establishment of temporal causality and understanding of usage trajectory changes.
• Experience effects undertheorized: While identified as important, experience was not formally modeled as factor modifying relationships.

Key Contributions

• Intention-behavior bridge: Shifted focus from measuring intentions to directly measuring actual utilization behavior.
• Social factor emphasis: Elevated social influence and organizational norms to primary status in technology adoption frameworks.
• Affective dimension inclusion: Included emotional attitudes as distinct from perceived usefulness, acknowledging emotional aspects of adoption.
• Organizational context integration: Situated model explicitly in organizational environments where job fit and organizational factors shape adoption.
• Multi-theoretical integration: Drew from Triandis, expectancy theory, and behavioral attitude theory, demonstrating theoretical eclecticism.
• Complex model structure: Specified indirect pathways and interaction effects exceeding simpler linear structures.

Internal Validity

The researchers established internal validity through rigorous quantitative methodology:

• Exploratory factor analysis: Identified underlying measurement dimensions, discovering seven factors rather than originally hypothesized six.
• Reliability assessment:Calculated Cronbach's alpha coefficients ranging from .60 to .86, with most scales exceeding .70 threshold.
• Discriminant validity: Confirmed that constructs loaded more highly on hypothesized factors than others, with appropriate construct separation.
• Path coefficient testing: Used jackknifing procedures not assuming normality to test significance, finding four of six hypothesized relationships significant at p less than .01.
• Sophisticated statistical techniques: Employed PLS analysis allowing simultaneous model testing with measurement error accounting.

External Validity

External validity testing was limited due to single-organization design, which the authors explicitly acknowledged:

• Diverse sample composition: Selected participants across different job levels and positions to capture variance in PC utilization across roles.
• Actual behavior measurement: Measured self-reported usage patterns rather than relying on intentions alone, strengthening external validity claims.
• Experience consideration: Collected data on previous PC experience and examined whether experience moderated model relationships.
• Adequate statistical power: Final sample of 212 knowledge workers across nine divisions provided sufficient power for detecting relationships.

Relevance to Technology Adoption

The PC Utilization Model is directly relevant to technology adoption because it identifies multiple barriers to usage and specifies organizational interventions reducing those barriers. By examining actual utilization rather than intentions, the model addresses practical organizational concerns.

Barriers to Technology Adoption Identified

• Perceived complexity: Users perceiving PCs as difficult to learn or operate show lower utilization regardless of system objective complexity.
• Poor job fit: When users perceive PCs do not align with actual job tasks or requirements, motivation to use them diminishes significantly.
• Negative long-term consequence perceptions: When employees perceive PC use will not lead to tangible benefits, career advancement, or productivity gains, utilization remains low.
• Unsupportive social context: Lack of peer support, organizational norms not supporting usage, or absence of respected champions undermines adoption.
• Insufficient facilitating conditions: Inadequate training opportunities, insufficient technical support, and poor resource accessibility create barriers.
• Weak positive affect: While affect did not show direct effects, emotional discomfort toward PCs may represent barriers in some contexts.
• Experience gaps: Inexperience with personal computers creates barriers through higher complexity perceptions and uncertain long-term consequences.

Leadership Actions the Model Prescribes

• Reduce complexity through training: Ensure accessible, context-relevant training reducing perception that PCs are difficult to learn and use.
• Establish clear job fit: Conduct job analysis identifying tasks where PC usage enhances performance and communicate connections explicitly to employees.
• Highlight long-term benefits: Educate on expected consequences including productivity gains, career development, and effectiveness improvements.
• Leverage organizational champions: Identify and empower early adopters and opinion leaders whose visibility publicly demonstrates adoption success.
• Provide comprehensive facilitating conditions: Ensure convenient physical access, responsive help desk support, flexible training, and documentation.
• Normalize technology affect: Destigmatize anxiety through positive experiences, highlight enjoyable computing aspects, and acknowledge comfort develops over time.
• Implement multi-level approaches: Target multiple intervention points simultaneously addressing complexity, job fit, consequences, social support, and facilitating conditions.

Following Models or Theories

The Thompson model directly influenced subsequent technology adoption research:

• Decomposed TPB (Taylor & Todd, 1995): Extended theory of planned behavior with technology-specific belief categories.
• Task-Technology Fit Model (Goodhue & Thompson, 1995): Built on PC utilization model by formalizing how technology characteristics align with task requirements.
• Unified Theory of Acceptance and Use of Technology (UTAUT): Incorporated multiple technology acceptance frameworks emphasizing actual usage.
• Technology Acceptance Model extensions:Subsequent TAM variants incorporated social influences and organizational factors similar to Thompson's approach.
• IS implementation research:Applied multi-factor frameworks similar to Thompson's model examining actual technology implementation outcomes.

References

1. Thompson, R. L., Higgins, C. A., & Howell, J. M. (1991). Toward a conceptual model of personal computing utilization. MIS Quarterly, 15(1), 125-143.

Further Reading

1. Triandis, H. C. (1980). Values, attitudes, and interpersonal behavior. In W. M. Alston (Ed.), Nebraska Symposium on Motivation (Vol. 27). University of Nebraska Press.
2. 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
3. Fishbein, M., & Ajzen, I. (1975). Belief, attitude, intention, and behavior: An introduction to theory and research. Addison-Wesley.
4. Porter, L. W., & Lawler, E. E. (1968). Managerial attitudes and performance. Irwin-Dorsey.
5. Goodhue, D. L., & Thompson, R. L. (1995). Task-technology fit and individual performance. MIS Quarterly, 19(2), 213-236. https://doi.org/10.2307/249689
6. Taylor, S., & Todd, P. A. (1995). Understanding Information Technology Usage: A Test of Competing Models. Information Systems Research, 6(2). https://doi.org/10.1287/isre.6.2.144

Series Navigation