Who are your peers?: An empirical investigation of the matched sample comparison analysis

Purpose – The purpose of this paper is to propose and evaluate a new matching sample comparison method, the industry size peer matching method. Design/methodology/approach – Based on archival financial data from Compustat and econometric methods, the paper first validates that such a method will result in firms being divided into more homogenous groups, making peer-performance comparison more meaningful. Then it compares this new peer matching method with previous methods through two resource-based related studies in the IT valuation context. Findings – The results show that the industry size matching method is a better method because: it is theoretically grounded, addressing industry, size, and random shock effects and, at the same time, avoids the selection bias caused by using a single firm as benchmark; and empirically such a technique results in more homogeneous groups and can explain more firm-level returns than the industry-only classification. Originality/value – Matched sample comparison group analysis is widely used in both academy and industry. The paper's theoretically grounds and empirically validated matching sample comparison method provides researchers and practitioners with a tool for their future research, performance evaluation, earning management detection, or compensation contract design, when selecting the right peers is called for.