THE RELATIVE EFFECTIVENESS OF HYPERTEXT AND TEXT

A series of two experiments was conducted. In Experiment 1, participan t performance when using a hypertext electronic reference system was c ompared to using a conventional reference book The links in this hyper text were based on the index entries in the corresponding 529-page boo k Specific topics and particular facts were located much faster and mo re accurately using the hypertext system than for the book. These adva ntages increased when participants searched for information that was e ither not included or referred to indirectly in the index. However, hy pertext did not have an advantage over text on learning tasks. The con clusion was that hypertext is superior to text only for ''reading-to-d o'' tasks similar to those a designer may perform when consulting a re ference book. Experiment 2 compared user performance when the links co rresponded exactly to the original index of a 545-page textbook on erg onomics to performance when the links were generated by computer key-w ord searches. Strong advantages were found in speed, accuracy, and sub jective ratings for links based on the author's original index. It app ears that these latter results can be attributed to the greater focus of the information provided by links based on the author's index. User s spent much less time browsing irrelevant sections of the book.