Experimental evaluation of sequential bias in online customer reviews

• We consider sequential bias, a type of bias that occurs in online reviews.
• Sequential bias is difficult to study using existing online retailer.
• This necessitates the use of controlled experiments.
• We experimentally study sequential bias in the context of online product reviews.
• Our results provide evidence for the existence of sequential bias.

Explicit information used in the development of recommender systems includes online customer feed-back. Such explicit input from customers suffers from several types of bias, which directly affect the quality of the resulting recommender system. Recent research has identified sequential bias to be present in online customer reviews. However, to our knowledge, no study to date has confirmed its existence in this context. Given the nature of sequential bias, confirming its existence necessitates a controlled experimental study, which is lacking in extant published literature. We attempt to address this gap. Results from our study show evidence for the existence of sequential bias.