Font Size: a A A

STRATEGIES IN SEARCHING HIERARCHICAL DATA STRUCTURES (INFORMATION RETRIEVAL)

Posted on:1987-03-09Degree:Ph.DType:Dissertation
University:The Ohio State UniversityCandidate:NORMORE, LORRAINE DOMBROWSKIFull Text:PDF
GTID:1478390017958717Subject:Psychology
Abstract/Summary:
This research investigated the way a data structure, the hierarchy, influenced the strategies people used in looking for a target embedded in that structure. Five search models were examined: random, branch exhaustive, branch sampling, level exhaustive and level sampling. A methodology was developed for identifying choice patterns characteristics of the search models. This methodology used uninterrupted sequences of consecutive choices ("runs") as the measurement unit. Subjects were given the task of finding random targets. However, their behavior was seldom categorized as random. Many choice sequences were most strongly influenced by the level dimension of the hierarchy. An interesting feature of the data was the tendency to use both level and branch dimensions to structure the strategies adopted. Subjects tended to use the physical arrangement of the hierarchy as a basis for systematic searching rather than the logical labelling. Two experimental manipulations were shown to influence subjects' tendency to search systematically. The behavior of subjects presented with systematically labelled data structures was more systematic than that produced by subjects presented with randomly labelled data structures. A second variation, referred to as aiding, had subjects cross off nodes as they were chosen. This permitted subjects to exclude previously chosen nodes from their choice set without reliance upon their memory for previous choices. It was shown that subjects who were not given aiding were more systematic (had a larger number of choices in the exhaustive category) than those using aiding.
Keywords/Search Tags:Data, Strategies, Search
Related items