Research On User Relevance Criteria Of Scientific Data

Scientific data has become a strategic resource for scientific researc.It is an urgent need for researchers to find the required data from the massive scientific data resources.Therefore,understanding the cognitive nature of relevance judgment of scientific data,providing theoretical basis for efficient discovery,quality judgment and governance of scientific data,has become a new subject of relevance research in the field of information science.User relevance research tries to understand the cognitive mechanism of information behavior,such as query and retrieval,so as to realize the interpretation and prediction of behavior.In terms of information carrier types,scholars have carried out exploratory studies on the judgment criteria.However,existing studies have shown that there are differences in information types of relevance criteria.Therefore,scientific data as a new type of information,how relevance criteria set and their impact on judgments will change is a question that needs to be answered.Based on the stimulus response theory and the lens model theory,the relevance judgement model was proposed.A total of 86 doctoral and master's degree students were selected in this study.The scientific data searchers' relevance judgment in real information needs was observed and recorded through the triangulation method including eye tracker record,in-depth interview and questionnaire.Using content analysis,analytic hierarchy process,statistical analysis and other methods to innovatively explore the scientific data user relevance judgment clues,criteria and their use.First of all,the study found nine scientific data relevance criteria,namely,topicality,availability,quality,standardization,authority,comprehensiveness,convenience,usability and currency.Topicality,availability,currency,authority,quality,convenience and usability are the common criteria across different information carriers.Comprehensiveness and standardization are unique criteria related to the characteristics of scientific data.And scientific data relevance standards have a certain level,which can be divided into five categories.Secondly,the research clearly distinguishes the scientific data's information attributions,clues,relevance criteria,and analyzes the corresponding relationship of the three.Information attributions were the external manifestations of data objectively existing in the environment.Clues were information that the judge perceived in the judgment.Judgment criteria were instrumental concepts in user cognition triggered by clues.Third,cognitive processing presents two characteristics: one stimulus,multiple responses;multiple stimuli,one response.Topicality was most complex because it required both common data information elements(such as title,abstract)and comprehensive professional knowledge.The rest criteria were relatively simple with little expertise.Fourthly,the study explored the characteristics of relevance criteria in users' long-term memory.The relevance criteria usage conformed to the long tail law and least effort principle.The whole judgment process conformed to the anchoring and adjustment strategy,the topicality was the anchor in the relevance judgment,then subsequently adjusted after the anchoring using other criteria.The judgment was the result of relevance criteria linear accumulation.Finally,scientific data relevance judgment model was put forward by integrating information attributions,clues,user relevance criteria and their internal relations.This study was innovative in the user relevance theory.The paper integrated the stimulus response theory and the lens theory,and put forward scientific data user relevance judgment theoretical model,which clearly showed the relevance judgment process of scientific data and enriched the theoretical research on the user relevance judgment.Secondly,the study comprehensively considered the properties and mutual relations of various variables in the relevance judgment process,and integrated eye movement observation,interview,questionnaire and other methods into triangulation method for the first time,realizing the innovation of the method.The future study will add more user type and situational type,to enrich the existing scientific data relevance judgment theory model,improve the description and interpretation ability of the model.