Institutionalism, Learning, and Patterns of Selective ...

Institutionalism, Learning, and Patterns of Selective Decoupling: The Case of TQM Daniel Z. Levin Rutgers University What Is Decoupling? Adoption X Implementation Many studies on adoption of innovation, where implementation is assumed, but Firms can claim to have adopted an innovation yet never fully implement it. (Meyer & Rowan 1977; Westphal & Zajac 1994, 1998)

Little quantitative research on the predictors of decoupling (Westphal & Zajac 2001) 2 Institutional vs. Learning Theory Question: Who decouples more: early or late adopters of an innovation like TQM? Institutional Theory: Late adopters, because they only adopted TQM to appear legitimate. (Scott 1987, 1995; Tolbert & Zucker 1983) Learning Theory: Early adopters, because they had few peers from which to learn how to implement TQM. (Argote 1993; Huber 1991) Which view is right? 3

Neither view is quite right. Selective Decoupling For complex innovations like TQM, implementation can occur along multiple dimensions. Selective Decoupling = More decoupling for some implementation dimensions, but less for others. 4 Two Dimensions of TQM Studied 1. Teams and Training Actively inspected by JCAHO (during site visits) beginning in 1995. 2. Use of Quality Tools

Only ceremonially inspected by JCAHO during this studys time frame. 5 Hypotheses Institutional Theory X Learning Theory H1: Late adopters will decouple less for teams and training H1: Late adopters will

decouple less for teams and training H2: Late adopters will decouple more for tools, H2-ALT: Late adopters will decouple less for tools Especially if overly burdensome; e.g., for advanced tools (H3) 6 Methods Used 1993 as dividing line between early vs. late adopters of TQM. (Reeves & Bednar 1993)

Two American Hospital Association (AHA) surveys of all community and VA hospitals: 1993: for early adopters of TQM vs. 1997: for late adopters of TQM 24 hospital interviews with managers, CEOs, doctors, nurses; and 2 interviews at JCAHO 7 Measures of Dependent Variables Teams and Training (4 items) % of FTEs on quality teams

% of doctors on quality teams % of FTEs with formal quality training % of doctors with formal quality training Use of Quality Tools (yes/no items) 6 basic tools 3 advanced tools 8 X DV=Teams and Training DV=Advanced Tool Use Late Adopters +.052

Late Adopters .053 Impl. experience Size Teaching Rural In system In alliance Competition + n.s. + n.s. n.s. n.s.

Impl. experience Size Teaching Rural In system In alliance Competition + + n.s. n.s. n.s. n.s. n.s. R2 = .138

R2 = .105 p < .001; p < .05 9 Conclusions Evidence of a subtle but predictable pattern of selective decoupling, depending on which implementation dimension one examines. Consistent with institutional theory, decoupling by late adopters decreases even below that of early adopters!for dimensions where late adopters are actively (as opposed to ceremonially) inspected. 10