Abstract
This article presents the findings of a replication study on the helpfulness of online reviews. The replication study utilized similar data and analysis procedures as the original study. However, the regression model reported in the original study could not be reproduced, forcing us to develop an entirely different model that had only three independent variables that could be verified as significant predictors of review helpfulness. In addition, the R 2 produced by the replication study was much smaller than the R 2 reported in the original study. Possible causes for the differences and implications of the new findings are discussed.