Looking for data (online survey)

Abstract

This survey dataset is part of the project "Looking for data: information seeking behaviour of survey data users", a study of secondary data users' information-seeking behaviour. The overall goal of this study was to create evidence of actual information practices of users
of one particular retrieval system for social science data in order to inform the development
of research data infrastructures that facilitate data sharing.

In the project, data were collected based on a mixed methods design. The research design included a qualitative study in the form of expert interviews and – building on the results found therein – a quantitative web survey of secondary survey data users. The survey dataset comprises 1,458 valid cases (1,727 cases including incomplete contributions). The transcripts of the expert interviews are also available through this data archive upon request.

The core result of this study is that community involvement plays a pivotal role in survey
data seeking. The analyses show that survey data communities are an important
determinant in survey data users' information seeking behaviour and that community
involvement facilitates data seeking and has the capacity of reducing problems or barriers.

In the quantitative part of the study, the following hypotheses were tested:
(1) The data seeking hypotheses:
(1a) When looking for data, information seeking through personal contact is used more
often than impersonal ways of information seeking.
(1b) Ways of information seeking (personal or impersonal) differ with experience.
(2) The experience hypotheses:
(2a) Experience is positively correlated with having ambitious goals.
(2b) Experience is positively correlated with having more advanced requirements for data.
(2c) Experience is positively correlated with having more specific problems with data.
(3) The community involvement hypothesis:
Experience is positively correlated with community involvement.
(4) The problem solving hypothesis:
Community involvement is positively correlated with problem solving strategies that require
personal interactions.

The calculations made to test these hypotheses can be reproduced with the syntax file LfdAnalysis.do that is provided together with the survey dataset.