Browse by author
Lookup NU author(s): Dr Kovila Coopamootoo, Professor Thomas GrossORCiD
This is the final published version of a report that has been published in its final definitive form by School of Computing Science, University of Newcastle upon Tyne, 2017.
For re-use rights please refer to the publisher's terms and conditions.
Abstract: Background. Scholars mostly agree that an attitude involves apositive or negative evaluation of a particular entity [2, 3, 4, 5], where theattitude object includes anything the person holds in mind. Evaluations canbe expressed via thoughts, feelings, intentions to behave and behavior.Together with cognitive responses (individuals’ belief and knowledge aboutan attitude object) and affective responses (individuals’ feelings and emotionsabout an attitude object), behavioral responses (the way the attitudeinfluences how individuals act or behave) form the three main classes ofresponses [6]. So far, privacy literature has not had a codebook for extractingprivacy or sharing attitude dimensions from individuals’ free form text.Aim. To provide a codebook for extracting privacy and sharing attitudes fromfree-form text.Method. Amazon Mechanical Turk participants were queried with “What does[privacy/sharing] online mean to you?” and asked to provide a responsewithin 250-words.First, an initial codebook was created from a sample of N = 18. We facilitateda conventional line-by-line coding of all response units, where each unit wasindependently coded by two coders. We obtained an initial codebook with aset of 43 concepts that are grouped across 7 categories.Second, the initial codebook was refined in several runs with 2 trained codersinto 6 categories and a total of 52 codes (Table 1, plus “Other” codes laterrefined into 29 finer codes (Table 2). We report on the detailed procedureand the full study in Coopamootoo & Groß [1]. The codebook was employedand inter-rater reliability computed, as reported in our privacy versus sharingprivacy attitude investigation [1].Results. The categories produced in the codebook and elicited for contentanalysis were the participant referring to: (a) himself (SEL), (b) who othersare in specific (SPE), (c) his emotions or moods (EMO), (d) others’ activities(ACO), (e) his own activities (ACS), (f) data or information (DAT). .We evaluate inter-rater reliability via %-agreement and Cohen k on 50 unitsacross the 52 codes. We find that the coders were on agreement 88.2% ofthe time. There was a substantial agreement between the two coders’judgment, k = .666, 95% CI [.630, .670], p < .001. We provide furtheragreement results across each category in Coopamootoo & Groß [1].Conclusions. This is a first codebook that enable extraction of cognitive,affective and behavioral dimensions of attitude, in particular designed forprivacy and sharing. We believe that it can be employed on privacy-relatedfree-form responses well beyond attitudes, such as to tease out userperceptions, emerging themes or mental models among others.
Author(s): Coopamootoo KPL, Gross T
Publication type: Report
Publication status: Published
Series Title: School of Computing Technical Report Series
Year: 2017
Pages: 13
Print publication date: 01/12/2017
Acceptance date: 01/01/1900
Report Number: 1518
Institution: School of Computing Science, University of Newcastle upon Tyne
Place Published: Newcastle upon Tyne
URL: http://www.ncl.ac.uk/media/wwwnclacuk/schoolofcomputingscience/files/trs/1518.pdf
