Research Ethics and Public Trust, Preconditions for Continued

Growth of Internet Mediated Research

Public Confidence in Internet Mediate Research

Ansgar Koene, Elvira Perez, Christopher J. Carter, Ramona Statache, Svenja Adolphs,

Claire O’Malley, Tom Rodden and Derek McAuley

HORIZON Digital Economy Research, University of Nottingham, Nottingham, United Kingdom

Keywords: Trust, Ethics, Privacy, Social Media Analysis, Transparency, Rights.

Abstract: In this paper we argue for the position that responsible safeguards for privacy and ethical treatment of

human data are of vital importance to retain the public confidence and trust that is necessary for the

development and future success of internet mediated research (IMR). We support our position based on the

high level of popular and media attention that is currently directed at IMR, which in combination with the

relative uncertainties that still exist around the ethics of various IMR methods, raises the risk that IMR

might succumb to a public backlash of similar proportions to the controversy that hit genetically modified

(GM) crops in Europe. Based on the lessons that came out of the GM crops controversy we discuss the

ethics requirements and challenges that must be met in order to retain the public trust in IMR. We end our

argument by briefly reviewing a couple of examples of “privacy protecting architectures” that are being

developed for IMR.

1 INTRODUCTION

The position we argue for in this paper is that

responsible safeguards for privacy and ethical

treatment of human data are of vital importance to

retain the public confidence and trust that is

necessary for the development and future success of

Internet Mediated Research (IMR).

Data derived from observations of human

behaviour and communication is intimate and

personal, no matter how or where the data was

collected. Furthermore, since such data is provided

by human participants who are ultimately free to

choose if they want to contribute to research studies,

or not, the ability to perform such IMR critically

depends on the level of trust that people have in the

IMR research community. If people understand how

their data is being used and can feel confident about

the benefits that the analysis of this data can offer to

themselves and society in general, then they will not

only willingly contribute their data but may even

choose to actively participate in further studies such

as citizen science projects.

The recent flood of media stories about leaks,

hacks and misuse of personal data is eroding

people’s trust in the concept of social media analysis

to the extent that they may soon rise up in a revolt

against all forms of IMR. Without transparency of

methods, clear ethics guidelines and technical

safeguards against (inadvertent) invasions of

privacy, public opinion could call for a boycott on

IMR similar to the backlash against genetically

manipulated crops that was triggered in the EU in

the 1990s (Carr and Levidow, 2000).

The argument for our position is structured as

follows: First we will set the scene by highlighting

the level to which IMR is currently “in the public

eye”. Next we draw upon the experiences from the

GM crops controversies in order to learn some

“lessons from history”. This is followed by a

discussion of the “Ethical requirements and

challenges for IMR” and a brief look at some

“privacy protecting architectures for IMR” that are

being developed. Finally we conclude by

summarizing the position of this paper and the

arguments we have presented.

163

Koene A., Perez E., J. Carter C., Statache R., Adolphs S., O’Malley C., Rodden T. and McAuley D..

Research Ethics and Public Trust, Preconditions for Continued Growth of Internet Mediated Research - Public Conﬁdence in Internet Mediate Research.

DOI: 10.5220/0005331501630168

In Proceedings of the 1st International Conference on Information Systems Security and Privacy (ICISSP-2015), pages 163-168

ISBN: 978-989-758-081-9

 2015 SCITEPRESS (Science and Technology Publications, Lda.)

2 IN THE PUBLIC EYE

Following in the wake of the explosion in popularity

and size of social media services over the last

decade, IMR including web-based questionnaires,

social media analysis and web analytics has rapidly

risen to become one of the most publically visible

forms of social science. For better or for worse, this

‘fame’ is for one part due to reflected glory from

association with flashy technology companies like

Google, Facebook and Twitter that are constantly in

the news, and for another part the result of

anticipated influence of quantitative data about

human behaviour generated by IMR on government

and corporate policy development. By producing

and accessing databases that are orders of magnitude

larger than traditional data sources (Dragland, 2013),

and doing so at speeds that can approach (near) real-

time processing of events, IMR is allowing new

types of quantitative and statistical analysis that are

promising to revolutionize social science. Both of

these developments have drawn the attention of

corporate, governmental and international

institutions interested in “evidence based policy”

(Wilsdon, 2014). Even the UN is looking to this data

revolution as an essential part of the global

development agenda after the 2015 millennium

development goals (MDGs) (Independent Expert

Advisory Group, 2014).

All of this fame and fortune (i.e. grant funding)

however carries with it a price tag in the form of

heightened public scrutiny, especially with regards

to the ethical conduct of the research (Booth, 2014;

BBC News, 2014) and possible failures to ensure

privacy protections (Barbaro and Zeller, 2006;

Narayanan and Shmatikov, 2009).

3 LESSONS FROM HISTORY:

GM CROPS

In this section we look at one of the more

controversial episodes in the recent history of the

relationship between science and the general public,

namely the European GM crops controversy of the

1990s. The case of the GM crops controversy is, in

our view, especially relevant since it also involved

research that, in the public eye, was strongly

associated with corporate interests and also strongly

hinged on the trust relationship between scientists

and the public.

In the 1990s the introduction of genetically

modified crops in Europe triggered a public backlash

against GM foods and biosciences in general (Carr

and Levidow, 2000). The cause of this controversy

has since been attributed to a combination of issues

relating to public trust in regulatory institutions,

scientists and industry (Frewer et al. 2004).

At the start of the 1990s, European regulators

discussed the issue of GM foods almost exclusively

as an issue of risks to the environment and human

health. The regulators failed to address people’s

fears that unintended effects are unpredictable and

thus unknown to science (Miles and Frewer, 2001)

or to adequately address the potentially

transformative nature of the technology on societal

and social structures, thus producing an erosion of

public trust in these institutions. This potentially

transformative impact of society is a concern we can

also find in the current scare stories about the ‘death

of privacy’ due to social media mining (Andrews,

2014; Jayson, 2014).

By failing to take due account of what was

driving public concern, the motives of those

developing the regulatory framework appeared

suspect. By appearing unconcerned with public and

environmental welfare the legitimacy of the

regulatory framework, and regulatory agencies with

responsibility for developing that framework, was

jeopardized. We do not believe that IMR has

reached this stage of public distrust yet.

The initial response to the public resistance to

GM crops was to dismiss many of the public fears

about GM crops as irrational and to attempt to gain

public acceptance by educating people with facts

and information. By the second half of the 1990s

this strategy had proven itself to be ineffective

(Biotechnology and the European Public Concerted

Action Group, 1998) and was replaced by a shift

towards participatory strategies. The main driver

behind these strategies was the belief that in order to

ensure societally accountable development of

technologies, e.g. genetically modified foods, there

must be a larger public involvement in debates and

decisions. Belief in this strategy was supported by

empirical research on the success of technological

innovations which repeatedly showed that early

involvement of end-users in the development

process significantly increase the likelihood of

market success (Cooper and Kleinschmidt, 1987;

Dwyer and Mellor, 1991; Wheelwright and Clark,

1992).

Ideally the participatory process proactively

engages in a debate with the citizens prior to the

development of technologies (and products) using a

variety of tools designed to engender active

involvement of the public. Such a debate not only

ICISSP2015-1stInternationalConferenceonInformationSystemsSecurityandPrivacy

164

reassures the public that their concerns are being

respected, but also produces a better understanding

of the public concerns, allowing them to be more

effectively introduced into risk assessment and risk

management practices.

In terms of lessons that could be learnt for the

introduction of new technologies in society, the GM

crops controversy clearly illustrated the importance

of maintaining the public trust, and the difficulties in

regain that trust once it has been lost.

The controversy also demonstrated how, in a

democratic society where choice exists, people

refuse to consume (literally or figuratively) what

they associate with some negative attribute.

In part as a response to the lasting aftereffects of

GM foods controversy on the biosciences in Europe,

the EU is now pushing for a much stronger emphasis

on Responsible Research and Innovation (RRI).

Under the RRI agenda (Sutcliffe, 2011) the aim is

to:

 deliberately focus research and innovation

to achieve a social or environmental benefit;

 have consistent, ongoing involvement of

society, from beginning to end of the

innovation process, including the public &

non-governmental groups, who are

themselves mindful of the public good;

 assess and effectively prioritise social,

ethical and environmental impacts, risks and

opportunities, both now and in the future,

alongside the technical and commercial;

 develop oversight mechanisms that are

better able to anticipate and manage

problems and opportunities and which are

also able to adapt and respond quickly to

changing knowledge and circumstances;

 make openness and transparency an integral

component of the research and innovation

process.

One element of this new approach to research and

innovation has been an increased interest in

participatory procedures to involve the public in the

decision making process in some way.

When considering how such participatory

procedures might be applied to IMR it is important

to acknowledge that social media users are likely to

have higher expectations about the level of personal

involvement in decision making than the general

population in the 1990s. For users who are used to

constantly expressing their opinions about the things

they find online, be it via comment boxes or ‘like’

buttons, participatory engagement with IMR policies

requires an ability to voice an opinion on the

perceived value (acceptability) of individual IMR

projects.

4 ETHICS REQUIREMENTS AND

CHALLENGES FOR IMR

Ethics guidelines and institutional review boards

play an important role in establishing an

environment of trust, where the public knows what

kind of research practices they can expect, and

researchers can gain confidence in their methods by

knowing who to turn to for an objective evaluation.

The Code of Human Research Ethics (British

Psychological Society, 2013) outlines four main

principles underpinning the ethical conduct of

research: 1. Respect for the autonomy and dignity of

persons; 2. Scientific value; 3. Social responsibility;

and 4. Maximising benefits and minimising harm.

For IMR the following issues often pose particular

challenges when evaluating the ethics of a

(proposed) study:

1. public-private domain distinction online;

2. confidentiality and security of online data;

3. procedures for obtaining valid consent;

4. procedures for ensuring withdrawal rights

and debriefing;

5. implications for scientific value and

potential harm.

4.1 Public-private Domain Distinction

When dealing with potential data derived from

online sources, e.g. discussions on user-groups or

social networks, the extent to which this should be

considered as being in the public or private domain

is often problematic. From a strictly legal

perspective only documents that are not protected by

public domain’. Regardless of the legal distinctions

between public and private interaction domains,

however, when confronted with data that was clearly

derived from specific online communications (e.g. a

Twitter post), the data will always be experienced as

intimate and personal by the person who posted it. In

order for people to feel comfortable with research on

such data therefore, requires that they have a high

level of trust in the researchers.

4.2 Confidentiality and Online Security

Anonymization is one of the most basic steps for

maintaining confidentiality, showing respect and

thus gaining the trust of research participants. The

ResearchEthicsandPublicTrust,PreconditionsforContinuedGrowthofInternetMediatedResearch-PublicConfidence

inInternetMediateResearch

165

need to protect the anonymity of participants is even

more pressing in IMR on social media posts where

access to the raw data, i.e. the on-line posts, cannot

be controlled by the researcher. At the same time,

the wealth of secondary information sources that can

be mined in connection to any hint at the identity of

a participant is making it increasingly easy to de-

anonymize research data. The classic example of

this is of course the de-anonymization of users in the

AOL Search Log by journalists of the New York

Times in 2006 (Barbaro and Zeller, 2006).

4.3 Valid Consent

Valid consent fundamentally deals with respect for

the autonomy and dignity of persons. In order for

valid consent to take place it is vital that the

participant is fully aware and has a true

understanding of that which is being consented to.

This is why, for instance, research involving

children requires consent from their legal guardian.

In many respects, the requirement for valid consent

represents a core value of any democratic society,

and yet it is probably the principle that is most

frequently violated on-line. An illustrative example

of this was the statement in the controversial 2014

Kramer et al. publication on IMR using Facebook, in

which they asserted that participants had provided

consent for the study since “it was consistent with

Facebook’s Data Use Policy, to which all users

agree prior to creating an account on Facebook,

constituting informed consent for this research”

(Kramer et al., 2014). The Data Use Policy however,

even if it was actually read by a Facebook user, does

not provide any information about the nature of the

specific Kramer et al study. The requirements for

full awareness and true understanding of the matter

that was being consented to was therefore clearly

violated in this study.

More in general however, any IMR study that

uses already available social network posts is

confronted with the problem of obtaining true valid

consent from the authors of those posts. While this

clearly presents a technical and potentially labour

intensive challenge, the advantage of consistently

making the extra effort of obtaining valid consent

will go a long way towards establishing a

conscientious and trustworthy reputation.

4.4 Withdrawal and Debriefing

The right to withdraw and the provision of adequate

debriefing are both closely linked to valid consent.

Since the act of participation often provides a deeper

understanding of the true nature of a study the right

to withdraw supports the validity of the consent

provided by the participants who remain in the

study. The same is true of the debriefing, especially

for research where the nature of the study requires

that the participants must be naïve to the true

purpose of the task/manipulations. One of the

challenges for IMR is the indirect or remote

interaction with participants which includes the

possibility of participants disappearing from the

study (e.g. closing a web browser page of an on-line

questionnaire) without communicating if they wish

their data to be removed and without paying

attention to debriefing information.

4.5 Scientific Value and Potential

Harm

Beyond issues relating to the execution of research,

a key element to evaluating the ethics of research

proposals is to establish if the study is likely to

produce results that are of scientific and/or social

value or if there are risks of potential harm. The

importance of dealing with these questions in a

thorough and transparent manner, and the potential

for a public backlash when this process is lacking,

was highlighted by the failure of the biosciences

community to clearly establish the rigor of their

ethics procedure in the case of GM foods. More

recently, controversies over social responsibility of

IMR studies are threatening to become a recurring

theme with stories like “Facebook reveals news feed

experiment to control emotions” (Booth, 2014) and

“OKCupid experiments with ‘bad’ dating matches”

(BBC News, 2014) being picked up in the popular

media. In order to counter this trend and retain

public confidence it is important to consider the

probable societal impact and response that a study is

likely to produce and weight this against the

anticipated scientific value.

5 PRIVACY PROTECTING

ARCHITECTURES FOR IMR

In order to encourage compliance with ethical

research protocols it is important to make this

compliance as easy as possible by providing tools

that are intrinsically privacy-respecting. This is

especially true for IMR where much of the research

is being done by corporations or other groups that

lack structures like the university ethical review

boards. Three such tools are:

ICISSP2015-1stInternationalConferenceonInformationSystemsSecurityandPrivacy

166

1. The PRISONER (Privacy-Respecting

Infras- tructure for Social Online Network

Experimental Research) architecture by

Hutton et al. (2012)

2. The Dataware system by Mortier et al.

(2013)

3. The Ma3tch (autonomous anonymous

analysis) technology.

PRISONER (Hutton et al, 2012) is an

architecture that was developed for conducting

social network experiments that preserve participant

privacy. The core element of the architecture is the

workflow manager unit that passes all data through a

data sanitiser before they are analysed or presented

to participants. The data sanitiser applies the

appropriate privacy-preserving transformations, that

are indicated by a privacy policy file.

Dataware (Mortier et al 2013) is a set of

technologies that were designed to enable people to

regain control over the digital data that is constantly

being created by and about them. Dataware provides

mechanisms for collating data that is held in

multiple locations (e.g. social media networks,

loyalty cards or banks) and making it available for

processing by third-parties, while retaining control

over the access to the data.

Ma3tch (Kroon, 2013) was originally built to

enable Financial Intelligence Units from various

countries to achieve virtual information integration

without infringing upon security, confidentiality,

privacy and/or data protection regulations. The

Ma3tch uses a ‘privacy by design’ framework that is

based on distributed agents to facilitate decentralized

but integrated information access, processing and

analysis. Relevant information and knowledge that is

distributed between autonomous organizations is

automatically detected and applied throughout the

network as soon as it emerges. Crucially the

sensitive raw data is never shared, only anonymized

standardized information is shared.

Each of these tools presents a different approach

to the problem of privacy preserving data handling,

and while much work is yet to be done, they do at

least provide examples to show that IMR can be

done without risking violations of privacy and

human dignity. As such, development of these tools,

and others like them, provides a clear signal that

there is no excuse for breaking the codes of ethical

research conduct. They also provide beacons of

research integrity to raise confidence and trust from

the public.

6 CONCLUSIONS

Based on the high level of popular and media

attention currently directed at anything related to

social media or the internet, IMR is currently

receiving a greater level of media scrutiny than most

other types of research. In combination with the

uncertainties that still exist around various aspects of

IMR ethics, this media attention carries the risk for

IMR of triggering a controversy and public backlash

similar to the one that hit GM crops in Europe in the

1990s. In order to avoid such a controversy it is

essential to retain the confidence and trust of the

public which, in the light of the “Snowden

revelations”, depends heavily on the use of

responsible safeguards for privacy and ethical

treatment of human data.

ACKNOWLEDGEMENTS

This work forms part of the CaSMa project

supported by ESRC grant ES/M00161X/1. For more

information about the CaSMa project, see

http://casma.wp.horizon.ac.uk/ .

REFERENCES

Andrews, L., 2013. I Know Who You Are and I Saw What

You Did: Social Networks and the Death of Privacy.

Free Press, New York, NY, USA.

Barbaro, M. and Zeller, T., 2006. A face is exposed for

AOL searcher no. 4417749, New York Times, August

9, 2006.

BBC News, 2014. OKCupid experiments with ‘bad’dating

matches. BBC News, 29 July 2014.

Biotechnology and the European Public Concerted Action

Group, 1998. Europe ambivalent on biotechnology.

Nature 387, 845–847.

Booth, R., 2014. Facebook reveals news feed experiment

to control emotions. The Guardian, Mon 30 June,

2014.

British Psychological Society, 2013. Ethics Guidelines for

Internet-mediated Research. INF206/1.2013.

Carr, S., Levidow, L., 2000. Exploring the Links Between

Science, Risk, Uncertainty, and Ethics in Regulatory

Controversies About Genetically Modified Crops.

Journal of Agricultural and Environmental Ethics, 12,

1, 29-39.

Cooper, R.G., Kleinschmidt, E.J., 1987. Success factors in

new pro- duct innovation. Industrial Marketing

Management 16, 215–233.

Dragland, Å., 2013; http://www.sintef.no/home/Press-

Room/Research-News/Big-Data--for-better-or-worse/

Dwyer, L., Mellor, R., 1991. New product process

ResearchEthicsandPublicTrust,PreconditionsforContinuedGrowthofInternetMediatedResearch-PublicConfidence

inInternetMediateResearch

167

activities and project outcomes. R&D Management 21,

31–42.

Frewer, L., Lassen, J., Kettlitz, B., Scholderer, J.,

Beekman, V., Berdal, K.G., 2004. Societal aspects of

genetically modified foods. Food and Chemical

Toxicology, 42, 1181-1193.

Hutton, L., Andrews, S., Henderson, T., 2012. An

architecture for ethical and privacy-sensitive social

network experiments. Proceedings of the ACM

SIGMETRICS. PADE’12.

Independent Expert Advisory Group, 2014. A World that

Counts, Mobilising the data revolution for sustainable

development. Report prepared at the request of the

UN Secretary-General by the IEAG on a Data

Revolution for Sustainable Development.

Jayson, S., 2014, Social media research raises privacy and

ethics issues. USA Today, March 12, 2014.

Kramer, A.D.I., Guillory, J.E., Hanckock, J.T., 2014.

Experimental evidence of massive-scale emotional

contagion through social networks. PNAS 111, 24,

8788-8790.

Kroon, U., 2013. Ma3tch: Privacy and Knowledge. IEEE

Interational Conference on Big Data. 23-31.

Miles, S., Frewer, L.J., 2001. Investigating specific

concerns about different food hazards—higher and

lower order attributes. Food Quality and Preference

12, 47–61.

Mortier, R., Lodge, T., Goulding, J., Skatova, A.,

Houghton, Crowcroft, J., Madhavapeddy A., 2013.

Dataware, A Personal Data Architecture. Digital

Economy 2013: Open Digital.

Narayanan, A., Shmatikov, V., 2009. De-anonymizing

social networks. Proc. IEEE Symposium on Security

and Privacy, May 2009. doi:10.1109/SP.2009.22.

Sutcliffe, H., 2011. A report on Responsibe Research &

Innovation. Matter. Obtained through the internet

http://ec.europa.eu.

Wheelwright, S.C., Clark, K.B., 1992. Competing through

development capability in a manufacturing-based

organization. Business Horizons 35, 29–43.

Wilsdon, J., 2014. Evidence-basaed Union? A new

alliance for science advice in Europe. The Guardian,

Mon 23 June, 2014.

ICISSP2015-1stInternationalConferenceonInformationSystemsSecurityandPrivacy

168