Page - 44 - in Radical Solutions and Open Science - An Open Approach to Boost Higher Education

46 D.Griffiths that it has been seen as constituting a ‘watershedmoment’ (McFarland,Lewis and Goldberg, 2015).The increasing influenceof ‘bigdata’ data-gathering strategies is blurring the linebetweenORandacademic social science in twoways. Firstly,OR is being applied infieldswhichwere once thepreserveof academic research.Until the emergence of the Internet, the data available toOR researchers was largely limited to that produced in-house by commercial and governmental organisations, such as production process data and internal communications. This led to awell-demarcated area of research,which studied the internal processes of organisations,andseldomclashedwithacademicresearchinvestigatingwidersocial phenomena. Today, however, huge quantities of data are available, generated by clientsandstakeholders’ engagementwithcomputer systems,whichenableorgani- sations toaddressamuchwider rangeofORquestions,manyofwhichare situated beyond the confines of the organisation per se. For example, retailerswould in the pasthaveconductedoperations researchon theirorganisationandcommunications, but would usually have relied on social scientists to analyze the social context in which theywere operating. Now, however, an advertisement for analytics roles at TESCOsays that successful applicantswill “help thebusiness to reallyunderstand ourcustomersandsuppliers” (TESCO,2019).Similarly, ‘predictivepolicing’prac- tices are encroachingon thedomainof academic criminology, usingdata analytics “to identify liktargets forpolice interventionandprevent crimeor solvepast crimes bymakingstatisticalpredictions”(Perry,2013p.xiii),combining,forexample,GPS tracking,licenseplatereaders,andgeographicprofilingtools(Perry,2013;Table5.3). Secondly, academic social sciencehas adoptedORmethods.According toGary King,Directorof theHarvard Institute forQuantitativeSocialScience, “Businesses nowpossessmoresocial-sciencedata thanacademicsdo” (Shaw,2014). In the face of this shift in power,many social scientists appear to have accepted that a change inresearchpractices is inevitable,withconcomitantethical implicationswhichhave yet to be spelt out. For example, the Social ScienceResearchCouncil has entered into the Social Science One collaboration with Facebook. This development has beenwelcomedbysomesocial scientists, forexamplePuschmann(2019),while the complaints of the EuropeanAdvisoryCommittee for ScienceOne (2019) are that “Facebookhasstillnotprovidedacademicswithanythingapproachingadequatedata access”. In contrast Leetaru (2019) argues that Social ScienceOnewill make the personal and intimate data of twobillionFacebookusers available for datamining by researchers, with little information available about the details of aggregation, privacyorhowtheresultsof the researchmightbeused in intervening insociety. In thecontextofdevelopments suchas this, it isunsurprising thatLApractitionersare unwillingtocommittotheconstraintswhichwouldbeplacedontheirrelativelysmall scalestudiesbycompliancewithpoliciesonresearchethicswithhumanparticipants. Theargumentmadehere is thatwithin thewider contextof the influenceofdata analytics on the social sciences, the development ofLAcanbe best understood as theextensionof theORresearch tradition to theeducation sector. Furthermore, the contradictions identifiedabovebetweenethical policies for academic researchwith humansubjects, andethicalguidelines for thepracticeofLA,correspondclosely to the tension between the ethical traditions ofORand academic research. I propose