Web search engines for children:: A comparative study and performance evaluation of Yahooligans!, Ask Jeeves for Kids™, and Super Snooper™

This study compared the search features and retrieval performance of Yahooligans!, Ask Jeeves for Kids(TM), and Super Snooper(TM) on identical searches. This extended the earlier research conducted by Bilal (1999), which examined the cognitive, affective, and physical behaviors of twenty-two 7(th) grade science students' use of Yahooligans! search engine. The three types of searches the students formulated in Yahooligans! on an imposed fact-driven query were here conducted in Ask Jeeves for Kids(TM) and Super Snooper(TM). The retrieval performance criteria of the three engines included: (1) retrieval output, (2) relevance, (3) overlap in results, and (4) redundancy. A matrix of features (e.g., database coverage, search interface, search capabilities, retrieval interface, filtering, feedback, online help, FAQs, and advertisements) was developed to compare the engines. The performance of the natural language capability embedded in Ask Jeeves for Kids(TM) was evaluated by comparing the results retrieved directly from Yahooligans! to those Ask Jeeves for Kids(TM) returned from Yahooligans! on identical searches. The results shed light on the strengths and weaknesses of each engine and appropriateness to specific types of queries. Yahooligans! was the most effective on keyword searching. Super Snooper(TM) returned the highest number of hits but with zero relevance on all types of searches. Ask Jeeves for Kids(TM) was successful only on one keyword search and failed both the natural language phrase and multiple keyword. Implications are made for improving the engines' design, retrieval performance, and search features, as well as for user instruction.