Building and testing a causal model of an information technology's impact

Abstract

An increasing number of firms are adopting Computerized Performance Monitoring and Control Systems (CPMCS) in an effort to improve the productivity of employees in labor-intensive service industries. The service sector has not historically used volume of output or similar quantitative measures of performance when evaluating employees. Thus, monitoring often represents a new evaluation method and a new application of information technology. It is an application prone to controversy: Proponents claim it improves measurement accuracy, fairness and consistency, while opponents argue that it degrades the quality of work life, increases stress and undermines customer sernce. Despite the need to understand the impact of CPMCS, there have been few attempts to predict what effects can be anticipated and explain how these effects arise. The methodology described in this paper was used to integrate existing anecdotal work with literature from reference disciplines to build a conceptual model of CPMCS impact on role definition. The three-phased research then used an intensive case study to build a theory of impact and generate testable, causal research hypotheses. Subsequent to this theory-building stage, surveys from 1500 service workers provided data to test the causal model. The research produced three outcomes. First, it combined theory building and theory testing in a study of information technology impact to give structure and direction to a field characterized by anecdotal research. Second, it provided two, tested causal models of CPMCS impact with good explanatory and predictive power. These models explained the influence of monitor design on attitudes toward production and customer service. Third, it demonstrated the use of a holdout technique to increase the amount of knowledge gained in the hypothesis testing stage of empirical research.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 44 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.