Partial profile · Book Profile — full narrative profile planned
Neural Networks in Organizational Research Applying Pattern Recognition to the Analysis of Organizational Behavior
A practical and conceptual guide to using artificial neural networks as pattern-recognition tools for theory development and applied problem solving in organizational and behavioral research.
A profile of this book is on the way.
What it’s about
Written for graduate students and practitioners familiar with multivariate statistics, this book demystifies artificial neural networks (ANNs) as a class of statistical procedures rather than mysterious black boxes or biological simulators. Scarborough and Somers argue that most organizational research rests on an unexamined assumption of linearity, and that neural networks—because they can model the full range of linear and nonlinear relationships among variables—offer a powerful way to revisit long-standing problems such as the job satisfaction–job performance relationship, employee selection, and organizational commitment. The book delivers the intellectual history of neural computing, a step-by-step guide to choosing software and training supervised (backpropagation) and unsupervised (self-organizing map) networks, two detailed real-world applications, and an honest accounting of the limitations and myths surrounding ANNs. Crucially, it insists that neural networks are a theory-development tool whose value depends on the researcher's judgment, not a shortcut that replaces careful science.
The through-line
- Who it’s for
- A graduate student, academic, or applied organizational researcher who is competent in multivariate statistics and wants to gain deeper, more accurate insights into human behavior in organizations.
- The problem
- Conventional linear statistical methods produce persistently weak effect sizes and low explained variance, failing to capture complex, nonlinear relationships in organizational data. The researcher feels frustrated, stuck, and worried that the field is stagnant, repetitive, or even subject to ridicule for its disappointing results.
- The plan
- Understand what neural networks are as a class of statistical procedures and their history.
- Learn when neural networks are appropriate versus when conventional statistics suffice.
- Choose suitable software and the right paradigm (supervised vs. unsupervised).
- Preprocess data carefully, then train, test, and evaluate networks while guarding against overtraining.
- Interpret network behavior using graphical analysis and sensitivity analysis, tied back to theory.
- The payoff
- The researcher uncovers meaningful nonlinear relationships previously masked by linear methods, opening new avenues of research. · Long-standing problems (like satisfaction–performance) are reframed and better understood. · Applied models (e.g., employee selection) improve prediction, tolerate noisy data, and support fair, consistent decisions.
See our guide
Related profiles we’ve built
- A meta-analysis of psychological empowerment: Antecedents, organizational outcomes, and moderating variables →
- A meta-analysis on social exchange relationships and employee innovation in teams →
- Antecedents of big data adoption in financial institutions →
- Identifying the Innovative Tension: A Meta-Analysis on Innovative Firm Behavior, Organizational Ambidexterity and Innovative Performance →
- Organizational interventions and occupational burnout: a meta-analysis with focus on exhaustion →
- The association between the interactive and diagnostic use of financial and non-financial performance measures with individual creativity: The mediating role of perceived fairness →
- The effect of human resource practices on psychological contracts at an iron ore mining company in South Africa →
- The Effects of Organizational Justice on Positive Organizational Behavior: Evidence from a Large-Sample Survey and a Situational Experiment →