Description:
Studies of online word of mouth have frequently posited---but never systematically conceptualized and explored---that the level of disagreement among existing product reviews can impact the volume and valence of future reviews. We develop a theoretical framework of disagreement in online WOM that, rather than considering disagreement as dispersion of opinions around a mean, argues that high levels of disagreement can be better conceptualized as opposing opinion poles. We use this framework to motivate, develop, validate, and test a novel statistical measure of disagreement, the Disagreement Index, that can be used to capture disagreement in a more nuanced manner than traditional dispersion measures such as standard deviation. We then demonstrate the use of the new measure by developing hypotheses about the effect of disagreement and test our predictions on nearly 300,000 reviews for 425 movies over three years. We find that only high levels of disagreement---with opposing opinion poles---influence future reviews while simple dispersion does not. We show that high levels of disagreement among previously posted reviews lead to more future reviews. We make the code for our tool publicly available to allow for ease of use in future studies of online WOM