Improving New Users’ Query Performance

Abstract

As the volume of data in organizational databases grows, organizations are seeking to use this data to improve organizational success. To this end, users are being asked to query these databases to provide information to help answer questions posed by key management personnel. Users who have had extensive experience with an organization's data can often detect the presence of errors in their queries when query results do not correspond to their ex ante expectations. New users, however, are less familiar with the data they will be querying. Having no, or limited, ex ante expectations for query results, new users may be unaware that the result produced by their query is incorrect. Unwarranted confidence in the correctness of their queries predisposes these users to stop looking for query errors even when their queries still contain errors. This behavior, premature stopping of query revision, prompts investigating whether new users' query performance would improve if they were not only provided with, but used, readily available information to form ex ante expectations. Our results demonstrated a threshold effect in new users heeding information for forming ex ante expectations. That is, the mere availability of information for forming ex ante expectations made no difference in query performance. When admonishing users to heed ex ante information, however, there was an associated increase in the accuracy of their queries. These results suggest that users unfamiliar with a particular database might make fewer query errors if they not only received readily available information but were then prompted to use the information to form ex ante expectations for query results. . 2012. Improving New Users' Query performance: Deterring premature stopping of query revision with information for forming ex ante expectations. ACM J.