Co-occurrence network analysis (CNA) as an alternative tool to assess survey-based research models in hospitality and tourism research

Abstract

Hospitality and tourism (H&T) researchers employ structural equation modeling (SEM) and other multivariate techniques to test their models with survey data. These approaches assess relationships among constructs and model fit, but they do not highlight the most influential survey items or links among them. Other challenges include method-specific requirements for appropriate data, the best indices to identify optimal models, minimum sample sizes, missing data, and interpreting the results from complex models. Co-occurrence network analysis (CNA) can mitigate these limitations. This study validates CNA in the H&T field with a survey dataset that assesses market strategy, nonmarket strategy (NMS), organizational values, and firm performance. CNA is proposed as a complement to existing multivariate approaches for assessing survey data. The assessment includes nine steps: (1) identify the research purpose and hypothesis, (2) determine the hypothesis-related items to measure, (3) determine the sample, (4) administer the survey, (5) determine the analysis method, (6) test the hypotheses, (7) prepare survey inputs for CNA, (8) employ CNA, and (9) visualize and interpret results. This pathway demonstrates how future research can apply and address CNA's advantages and limitations.