LAUSR

Purpose This paper aims to examine how sample design affects the relative importance of firm and industry factors in explaining performance variations. Design/methodology/approach Using a sample of 14,204 Spanish firms over a 10-year time frame, this study uses partial sensitivity analysis to examine the biases in results as a consequence of three methodological relevant concerns: outliers, industry classification and period. Findings Results indicate that the industry effect, supported by the industrial organization theory, has been underestimated in the empirical tests. Originality/value This study examines the biases in results as a consequence of three methodological relevant concerns (outliers, sector classification and period), which have not been sufficiently studied to date. Moreover, the study provides some new evidence favourable to the Industrial Organization (IO) perspective, which could have been biased and underestimated by the literature, as most of the analyses do not consider the methodological issues studied in this paper.