| Chapter 1: | Introduction |
It considers looking at banner ads to be a secondary task. Assuming that users have a limited capacity of mental resources, processing banner ads should interfere with a primary task when both tasks require more resources than are available in users’ information processing systems. Thus, the studies using this method measure performance on the primary task, and deterioration in this performance is considered to be an indication of both processing of the banner ads and interference with the primary task. A study by Burke and Hornof (2001) provides a good example of using the interference paradigm. This study examined whether the presence of banner ads during a visual search task would decrease performance on the task. In this study, subjects were instructed to find a particular target and click on it with the mouse. The target was a four-letter word and was presented before each search trial. In each trial, there was one target and 19 similar distractors. Two banner ads were placed in random rows, both of which were either nonanimated or animated banner ads. Targets were placed in random positions across trials, and search time for the target was measured. It was predicted that the interference produced by animated banner ads would slow the search time compared with nonanimated banner ads. This prediction was not supported, suggesting that no interference had occurred. However, they also found that recognition of banner ads was higher for nonanimated banner ads than for animated ads.
What this example demonstrates is that even though a slower search time is interpreted as indicating greater interference from banner ads, the lack of a decrement in search time does not prove that processing has not occurred. In addition, if search time is slowed, the nature of the interference is not known. Right now little evidence exists to indicate that more interfering banner ads can be equated with greater processing of the banner ads. For example, Hong, Thong, and Tam (2004) found that search time for targets was slower when distractors were flashed, but recall of the flashed distractors was no better than that of nonflashed distractors.