David Knoke

Log-linear models

Discusses the innovative log-linear model of statistical analysis. This model makes no distinction between independent and dependent variables, but is used to examine relationships among categoric variables by analyzing expected cell frequencies.

Strategic Alliances and Models of Collaboration

The purpose of this paper is to engage in a comprehensive review of the research on strategic alliances in the last decade. After presenting a typology of diverse alliance governance forms, we review recent analyses of alliance formation, implementation management, and performance outcomes of collaborative activities. Our findings reveal that strategic alliances developed and propagated as formalized interorganizational relationships. These cooperative arrangements represent new organizational formation that seeks to achieve organizational objectives better through collaboration than through competition. The paper provides future research directions on partner selection, networks patterns and processes, understanding the integration in alliances through fusion, fission, and how to manage developmental dynamics. We conclude with some future directions for theory construction and empirical research.

How Organizational Field Networks Shape Interorganizational Tie-Formation Rates

We investigate the impact of communication in field-level networks on rates of formation of interorganizational collaborative ties, such as strategic alliances and joint ventures. After developing the concept of an organizational field network (“field-net”), we derive a set of testable propositions and corollaries that relate field-net properties, such as density, reciprocity, centralization, multiplexity, and hierarchy, to subsequent nonlinear changes in interorganizational tie-formation rates. We conclude by discussing aspects of empirical research for testing the empirical validity of these propositions.

A Comparison of Log-Linear and Regression Models for Systems of Dichotomous Variables

The relative abilities of dummy variable regression and log-linear models to locate significant relationships in systems of dichotomous variables are compared. On logical grounds log-linear models are superior to regression since the data more readily meet the assumptions of the former. Two illustrative examples suggest that the methods converge in their findings when the range in proportions of the dependent dichotomy is between .25 and. 75, but may differ on which effects are significant when proportions are more extreme. Substantive differences under the two methods are likely to be small, however.

JOB TRAINING IN U.S. ORGANIZATIONS

We draw hypotheses about the factors related to company provision offormal job training programs from diverse theoretical perspectives and research findings. Using data from the 1991 National Organizations Survey, we analyze a sample of 688 establishments to estimate multivariate models. The relationships between employer-provided job training and organizational size, unionization, and workforce composition are reduced or eliminated in most multivariate equations. Employer-provided training is most extensive in establishments with elaborate internal structures that operate in complex market environments. The job-training practices of U.S. employers affect employees from the executive suite to the loading dock, enhancing skills from basic literacy to interpersonal sensitivity. A new training ideology is rapidly eclipsing the traditional segmented pattern-a pattern in which universities educate the professionals and technicians, companies prime the executives, unions apprentice workers in the skilled trades, and government prepares the disadvantaged. Facing competitive world economic pressures that have eroded Americas market positions, employers are now using job training as one means of coping with changes fostered by technological innovation, market competition, organizational restructuring, and demographic shifts. Although training is an integral part of the employer-employee relationship, direct evidence about company training practices based on representative samples of diverse employing organizations is almost nonexistent. Most previous research has drawn on either self-reported labor force surveys or highly restricted samples of organizations (typically skewed to very large firms). We examine hypotheses about employer-provided job-training practices in the early 1990s, using a national survey of U.S. work organizations.

The Spread of Municipal Reform: Temporal, Spatial, and Social Dynamics

The adoption of commission and managerial administrative structures and their subsequent abandonment by the 267 largest American cities during the period 1900-1942 is studied using event-history methods. Full-period equations show no significant direct effect of percentage Catholic (culture clash) or population size (hierarchical diffussion) and only a modest impact of city age (modernization). A strong effect from average annual manufacturing wages is found, but in a direction contrary to that posited by the class-conflict hypothesis. The variable affecting transition rates most strongly appears to be regional adoption percentages, an indicator of neighborhood diffusion. These findings suggest revisions of the conventional image of the municipal reform experience.

Commitment and Detachment in Voluntary Associations

Membership commitment to and detachment from voluntary associations is hypothesized to be affected by the degree to which the organizations political structure facilitates social control by members, in the sense of mutual self-regulation. Three components of the polity-decentralization, communication, and total influence-are analyzedfor their impact on members of 32 noneconomic voluntary associations in Indianapolis. Using a contextual-effects model, communication and participation in decisions are found to increase commitment and to reduce detachment net of other organizational and individual factors. An interaction between communication and decision participation suggests that communication can compensate for lack of direct involvement in making decisions in strengthening member support for the collectivity.

Beyond Profit? Sectoral Differences in High-Performance Work Practices

Drawing on a recent survey of establishments in the United States, the authors examine how nonprofit, public, and for-profit establishments vary in the use of high-performance work organization (HPWO) practices that offer opportunities for participation in decision making (via self-directed teams and offline committees), enhance the capacity for participation (via multiskilling practices such as job rotation), and provide incentives for participation (via compensation practices). Nonprofit and public organizations are less likely to use performance incentives (gain sharing and bonuses) and some multiskilling practices than are for-profit organizations but more likely to use both self-directed work teams and offline committees. Sectoral differences in the prevalence of incentive compensation and self-directed teams persist after correlates of sector that predict HPWO prevalence—including establishment size, industry, computational requirements, and unionization—are controlled.

The Gender Gap in Company Job Training.

Previous analyses of employee participation in company-provided job training programs, using cross-sectional data on cumulative incidence levels, found either that men receive more training than women or no significant gender differences. The authors conducted event-history analyses of the hazard rate of entry into initial firm training programs by a national cohort of young workers. Rather than closing the gender gap, the womens training disadvantage widened after controlling for theoretically important human capital, occupational, industrial, organizational, and family-stage variables. Further examination of womens and mens distributions on these independent variables and estimates of separate event-history equations suggest that gender segregation by occupation and industry, workweek length, and family role obligations afford men better training opportunities than women. The authors conclude with suggestions for future research and speculations about the policy implications for closing the persistent gender gap in company-provided job training.

Measuring Organizational Structures and Environments

This article provides a descriptive overview of the work establishments in the National Organizations Study (NOS). It begins by reviewing their auspices, industry settings, and composition. Next, it introduces the survey items and scales used to measure coordination and control structures including structural differentiation, formalization, decentralization, and the presence of internal labor markets. Other items and scales refer to aspects of the technical and institutional environments. The NOS includes a quite diverse set of establishments, most of which provide services rather than produce goods. The public, nonprofit, and private, for-profit sectors are well represented. A sizable fraction of the workplaces are parts of larger organizations. Differences between descriptive statistics for the unweighted and weighted NOS samples highlight differences in the structures and environments of larger and smaller establishments.