In this chapter, we investigate several methodological challenges inherent in survey-based analyses of the impact of high-performance work systems on firm performance. Drawing on a national panel survey of organizational human resource (HR) management systems, we compare the estimated relationship between HR strategy and firm performance in both cross-sectional and longitudinal data sets. Prior research relying on multifirm data sets has typically relied on cross-sectional estimates that are potentially subject to problems of unobserved firm-level characteristics, such as the quality of marketing or manufacturing strategies, that might bias the estimated HR strategy-firm performance relationship. While panel data can mitigate such heterogeneity bias, such data are even more sensitive to the attenuating effects of error in the measurement of HR management practices. Thus the main objective of this chapter is to provide direct estimates of the likely magnitude of both heterogeneity bias and measurement error in cross-sectional estimates of the HR strategy-firm performance relationship.

The potential strategic impact of high-performance work systems is consistent with a new focus in the literature on behavioral strategies that rely on core competencies and capabilities as sources of competitive advantage, not only because they provide the most effective response to market demands, but also because they are not easily copied by competitors (Prahalad & Hamel, 1990; Stalk, Evans, & Shulman, 1992). A key element in the implementation of such strategies is the extent to which a firm's HR strategy, as reflected in the adoption of a high-performance work system, supports these larger strategic objectives (Huselid, 1995).