Abstract
The authors propose a concept of “search disclosure” to assist in the understanding of the willingness of a user to moderate their digital information seeking behaviour as a result of the perceived anonymity afforded at the point of searching and information consumption. Differences in the way people search for health information on touch‐screen kiosks, the Internet and digital interactive television are thought to result partly from the anonymity of the place in which the search is conducted. Data have been drawn from 11 independent studies, involving questionnaire, interview and log analysis methods across the three digital information platforms. These studies were all funded by the Department of Health as part of the evaluation of the national roll‐out of digital information and advice services to the consumer. Search disclosure is important in not only providing an understanding of existing consumer search behaviour but also in helping to design personalised online services, such as the electronic patient record. The data presented are speculative in nature and further work is being conducted to gather more robust data.
Keywords
Citation
Nicholas, D., Huntington, P., Williams, P. and Gunter, B. (2003), "“Search disclosure”: Understanding digital information platform preference and location in a health environment", Journal of Documentation, Vol. 59 No. 5, pp. 523-539. https://doi.org/10.1108/00220410310499573
Publisher
:MCB UP Ltd
Copyright © 2003, MCB UP Limited
Introduction
Research at City University, funded by the Department of Health, comparing different digital platforms (touch‐screen kiosks, Internet and digital interactive television) in respect to their value as sources of health information and advice has highlighted a particular characteristic of information seeking behaviour. Whether and how a particular digital information platform in a certain location is used appeared to depend to some extent on the “privacy” that the platform and environment offered. For the purposes of this discussion we have called the willingness or otherwise of people to search for information on a given digital platform and/or in a particular location – “search disclosure”. As a concept, it has similarities with the notion of self‐disclosure used in social psychology and psychotherapy. Self‐disclosure refers to the extent to which an individual will reveal personal details about him/herself to another. Such disclosure may vary with the nature of the communication channel (Joinson, 1998; Joinson and Banyard, 2002). We are particularly interested in how we can use the concept for understanding user interaction in the digital environment, especially with regard to the health field where search disclosure would seem to be an important issue. The anonymity offered by certain digital health platforms in certain locations might be seen as an asset when communicating about highly sensitive topics, such as sexually transmitted diseases, or embarrassing topics, like incontinence.
The motivation for developing the idea of search disclosure first arose from previous research by the authors (Nicholas et al., 2002a), which found considerable differences in use between health information kiosks that were located in different places. The research found that health touch‐screen kiosks under‐performed where kiosks were located in situations where the user was required to search in the presence of others, for example, in a seated waiting room area. This was thought to inhibit IT use in this context because users might feel uncomfortable disclosing their need for medical information.
It would seem logical that the willingness of a person to search a digital database for all or certain topics is mediated by the perceived anonymity afforded by the system. The concept is tied to issues of anonymity and sensitivity about revealing possibly intimate information needs. Search disclosure tells us something about how the user might be discouraged or encouraged to use a particular information system in a particular location for a particular topic.
Previous literature
The general concept of search disclosure (regardless of the actual terms used) appears to be a new one in information science. Some – but not much – research has been carried out into non‐use or low use of information systems. Julien (1999), for example, points out that people's affective (emotional) states lead them to seek information from human sources rather than from “systems” and, for example, Wilson (1995) also considers that oral communication is a fundamental part of certain types of information seeking. Finally, and most relevantly, Nahl and Tenopir (1996) suggest a concept of “search stress”, which orients information searchers towards human sources when seeking answers to questions that offer confirmation and reassurance. In the health field, studies tend to concentrate on varying information needs and information avoidance per se, rather than preferences for certain delivery mechanisms. Leydon et al. (2000), for example, showed that, beyond routine information about treatment, many patients would “rather not know” specific information about their conditions. These authors concluded that “cancer patients’ attitudes to cancer and their strategies for coping with their illness can constrain their wish for information and their efforts to obtain it”. Earlier research with people with cancer came to similar conclusions (McIntosh, 1977).
In psychology, much work has been undertaken on “self (as opposed to ‘search’) disclosure”. This research has examined the willingness of people to disclose information about themselves in person‐to‐person communication, under certain conditions – principally depending on the medium employed. As Joinson and Banyard (2002) point out, the psychological processes inherent in information seeking on the Internet, rather than on person‐to‐person communication, have received little attention, with most studies focusing instead on evaluating content or search strategies. Joinson (1999), for example, compared online and hardcopy questionnaire responses, and found that participants who communicated online were more honest about themselves. Coleman et al. (1999) found greater self‐dislosure in computer‐mediated correspondence as compared to face‐to‐face interactions.
“Search disclosure” model
We envisage a variety of location and platform opportunities where users might take up broadly similar types of information content and assume that “search disclosure” will have a varying impact in each case. Search disclosure is all about people being sensitive to providing information to a particular digital platform in a particular place because of their perceptions of how anonymous or not they were. Environments and platforms offering less anonymity will be less well used for some sensitive topics. Thus, under this model, touch‐screen kiosks will generally be less used to search for some health topics than are an Internet or digital interactive television (DiTV) health information system because touch‐screen kiosks in general offer less anonymity to the user.
Each particular platform (touch‐screen kiosks, Internet and DiTV) offers potential users a variety of search situations. In general, though, kiosks are usually positioned in public places, albeit offering various levels of anonymity. However, there are a few examples of kiosks being located in a dedicated room where users can search for information in privacy. Although in theory, the Internet and DiTV, especially the latter, offer much more privacy, there are examples where they offer no more privacy than the kiosk. Privacy may be diminished, for instance, with a shared access computer in an office environment or with a TV set in a sitting room where others are present or might appear without warning at any point. Furthermore, a kiosk based in a busy hospital serving a large catchment area might offer users greater anonymity than a kiosk based in a GP's surgery, where you might know someone or where you are obviously engaged in an activity when most people are just biding their time, looking around etc. A bedside TV set might also offer the opportunity for relatively private use in the case of hospitalised patients.
Aims
This paper retrospectively explores evidence for the existence of search disclosure in a health setting. Typically this can be stated in the following conjecture:
The greater the users’ perceived need for search privacy, the more likely it is that that search will be done when privacy is more assured.
Given knowledge of search environments the conjecture can be broken down into additional conjectures given the users’ perceived need for search privacy. The following hypotheses are considered: that searching at home is preferable to searching in a public place (be it office or public); a public environment, where the searcher is less likely to be observed, is preferred to public places where the users would more likely to be observed. For example, a searcher might choose a particular time or location to search on a sensitive health topic.
By examining interview, transaction log and questionnaire data gleaned from a range of digital health information platforms collected as part of a wider study on the impact of these electronic health information platforms on the health consumer. It seeks to demonstrate the possible impact of search disclosure in a number of situations (information platforms and locations) offering different levels of privacy and anonymity. The hypothesis being tested is that the greater the perceived need for privacy, the more likely it is that searching will be done where and when privacy is more assured.
While we present a wealth of evidence to support our supposition, the article has to be speculative in nature as in none of these investigations did we set out specifically to gather information on search disclosure (as we were not even certain then that it existed). Instead, the concept was formulated to explain differences in patterns of use emerging from many of our studies. The results could be confounded by additional variables; however, extensive research comparing platforms (Nicholas et al., 2003b) has not revealed a more plausible model, Further the supporting evidence draws on a number of independent studies and hence minimises the risk of a bias a confounded result. Future research will reveal if search sensitivity to the same information need varies by gender, age, social class or ethnicity. The objective of the paper is then to shed further light on the phenomenon and point to areas of future research, which we shall progress through our continuing research commitment to the Department of Health.
Methods
This paper draws on eleven qualitative and quantitative studies that we have conducted as part of a major investigation of the impact of a range of digital health consumer platforms, which suggest that search disclosure might be impacting on information seeking behaviour. However, it must be acknowledged that it is difficult in all cases to attribute differences in searching behaviour exclusively to search disclosure, as other factors might come into play. Thus, there are also issues of search comfort, perceived ease of use, digital visibility and user system expertise that also emerged as significant mediators of online behaviour in our earlier work. These factors cannot be dismissed and this is especially true when looking at patterns of online behaviour revealed by log files. Brief details of these studies follow and, for more details, the references provided in the bibliography should be consulted:
- 1.
Qualitative studies:
- •
In‐depth interviews with 25 users and non‐users of Intouch with Health touch‐screen kiosks, in various medical (doctors’ surgeries) and non‐medical (pharmacies) locations (Williams et al., 2003).
- •
An online “open‐ended” survey hosted on the NHS Direct Online Web site for three weeks during November 2002. A total of 42 people responded. The questionnaire asked why respondents chose the Internet for health information, and the kind of information being sought (Nicholas et al., 2003b).
- •
In‐depth interviews, 31 respondents, of Birmingham City Council employees use of a broadband health workplace information service delivered to workplace computers (Nicholas et al., 2002b).
- •
In‐depth interviews with 21 subscribers to Living Health, a digital TV health information channel (Nicholas et al., 2002b).
- 1.
Questionnaire studies:
- •
An online questionnaire hosted on the NHS Direct Online Web site. Responses collected over 13 months from the last week of 1999 to January 2001, 3,374 users replied – less than 0.5 per cent of users using NHS Direct over this period. The questionnaire asked users about their site visit, facilities used and how they found the site (Nicholas et al., 2002b).
- •
A postal questionnaire sent out in July 2002 of patients (aged over 16) of a GP surgery (Esk Medical Practice) that hosted an InTouch with Health touch‐screen information health kiosk located in a corner to the side of the main reception area. A random sample was selected and 190 respondents (patients) replied (approximately 8 per cent of all patients) (Nicholas et al., 2003a).
- •
A postal questionnaire sent out September 2001 to “all” subscribers to Communicopia, a digital TV health information channel. About 12 per cent of the population, 1,184 people, responded, and surveyed as to their use of this digital information services (Nicholas et al., 2002b).
- 1.
Web log studies:
- •
A transaction log analysis of 21 Intouch with Health kiosks for the period March 1999 to April 2001. During the period of the study 88,525 user sessions were recorded (Nicholas et al., 2001).
- •
An analysis of SurgeryDoor health Web site[1] log files for the 12 month period October 2001 to September 2002. During this period the site was visited by 381,704 separate IP addresses (excluding declared robots) (Nicholas et al., 2002b).
- •
NHS Direct Online (NHS DO) Web logs for November 2000. During this period NHS DO received 41,510 individual visitors (separate identifiable IP addresses) (Nicholas et al., 2002b).
- •
An evaluation of the logs of the Living Health consumer health DiTV channel for the period 22 August 2001 to 17 September 2001[1]. During this period the site attracted 6,903 individual visitors (Huntington et al., 2002).
Results
Touch‐screen kiosks
We set out an argument to show that kiosk use is moderated negatively by search disclosure. That is, people experience discomfort using IT platforms situated in public areas when searching for private, confidential or embarrassing information. This factor may be especially likely to influence the use of a digital health platform such as a touch‐screen kiosk in a pharmacy or doctor's surgery, where other people may be physically present.
Thus, in a comparison between kiosk use at four location types (information centres, pharmacies, hospitals and surgeries) pharmacies performed poorly (Nicholas et al., 2002a). Pharmacies recorded the lowest number of pages viewed in a search session, under six page views on average compared to about seven at other locations. Furthermore, they recorded the lowest average session duration time of less than 50 seconds. Kiosks located in hospitals and information centres recorded the longest average session duration time of approximately 80 seconds while surgeries recorded an average session view time of about one minute. Surgeries and pharmacies recorded the shortest average page view time of about eight seconds while users at hospitals and information centres recorded average view times of about ten seconds. In terms of overall use, estimated by the number of sessions per hour per kiosk, kiosks located in information centres performed well and recorded just over one user session per hour. Surgery kiosks were the most under‐deployed and recorded on average around one session every two hours. Pharmacies attracted just under two sessions every three hours, while hospitals performed slightly better than this and recorded just over four sessions every five hours. Surgeries also performed poorly according to the number of pages printed per hour per kiosk and surgery locations recorded just under one page printed per day. Information centres again performed well with one page printed every two‐and‐a‐half hours. Pharmacies performed poorly with less than one page printed every four hours.
Surgeries appear to perform the least well – especially according to the number of sessions conducted in an hour and the number of pages printed per hour. This may not be surprising. After all, surgeries are not the most relaxing of locations, and people are faced with time constraints and doubts before their appointments, not knowing how long they will have to wait beforehand. In fact, patients may be eager to get their appointment over with. Also, locations with a large throughput of people, such as hospitals, will record higher levels of use. However, some of the variation in kiosk use between locations is thought to result from the anonymity offered by the location, with surgeries and pharmacies being the most visible of all locations and surgeries being the most personal.
Data obtained from a questionnaire study on people who had not used the kiosk supports this belief. Evidence of search disclosure was uncovered in the questionnaire study of potential users of a kiosk situated in a doctor's surgery in Scotland. Respondents who did not use the kiosk were asked whether the reason for non‐use was anything to do with the fact that the kiosk was situated in a public place. Just under half of non‐users (47 per cent) said that they did not like the idea of using the kiosk because it was in a public place. There was some evidence that older users were more likely to say that they did not like using the kiosk in public place: 56 per cent of over 55 years agreed compared to 32 per cent of non‐users aged 35 and under. The search‐disclosure model predicts that potential users would be put off using the kiosk if issues around privacy were compromised. In fact, a little less than 50 per cent of those who were identified as non‐kiosk users preferred not to use the system rather than conduct their information search in a public place.
This was further confirmed from interview data conducted in pharmacies, in which various issues related to kiosk use were explored. Although interviewees were not asked specifically about the positioning of the kiosk in relation to their propensity to use it, some respondents stated that use of the system for them would be dependent on the actual placement at the location. One user commented that “If it is positioned in such a way that someone can see over your shoulder I would not want to use it”, while another said that “yes, I would use it anywhere as long as it is in an area that offers some amount of privacy”. One interviewee even said “there is always the chance of running into someone I know there, I would feel uncomfortable for them to see me searching for information in a doctor's surgery”. Finally, one respondent highlighted the hazards of going ahead with a search interaction: “I was trying to print some information and the machine stuck. I asked the pharmacist for help even though I was a bit embarrassed because I considered that particular information rather confidential, but I really needed it. She assisted me in printing all my pages”. These are issues of search disclosure in that users were fearful of their search being seen.
Relative topic use at the various kiosk locations also provides some evidence of self‐disclosure. Thus, in comparing the use of the kiosk page on depression, a clearly sensitive topic, between four kiosk locations – hospital, pharmacy, information centre and surgery – it was found that users were about 25 per cent more likely to search for this topic at a hospital compared to either a surgery or an information centre.
Internet
Searching the Internet may offer users’ some anonymity, albeit depending on the location of the terminal. Typically, use tends to be in environments where the user has their own personal space, where there is likely to be a degree of privacy and on someone's own machine. Clearly, the degree of privacy associated with PC‐based Internet behaviour can vary. Use of a PC to surf the Web in an open‐plan office may not offer the level of search privacy that accessing the Web via a PC at home does.
A study was conducted that used Web logs to compare the use of depression (a possibly sensitive topic) and healthy eating (a less personal/sensitive topic) pages by hour of day. The aim was to find out if these potential differences in content sensitivity manifested itself in use patterns. The analysis was restricted to UK users only so as to restrict the analysis to a single time zone and to commercial registered domain names. Unfortunately, these domains are also used by IP providers and thus this group does not consist wholly of commercial organisations but will include home users linking to the Internet via IP providers with a “dot co” address; for example, www.demon.co.uk We have stated in the methods section that search disclosure is one of many effects and is difficult to isolate. In fact, the diurnal patterns may well reflect other factors associated with these two groups of users: users of depression pages and users of healthy eating pages. A similar analysis was repeated with DiTV (Figure 1) giving perhaps a wider insight into diurnal patterns between these platforms as well as within platforms. If there are specific influences related to use of depression pages other than search disclosure and wanting to find a “quiet” time to search this topic, then this should show up in a comparison between the Internet and DiTV. However, it does not look as if there is. The diurnal patterns of online searching for information about depression follow the same general trend in each case but seem to be accentuated at certain periods of the day.
Figure 2 compares use of pages on depression and healthy eating on the SurgeryDoor health Web site over the course of a day, divided by hour of day. It was posited that depression was a more sensitive search subject and that some users at least would seek a more private period to search for this information. Figure 2 compares the percentage share of use over hour of day for these two topics.
There was a tendency for a greater use of depression pages at periods that could be described as more private: in the evening from about 6.00pm to about 2.00am and at about 1.00pm as people go out to lunch. Users may be waiting till after 6.00pm when again the office may generally be less busy. The use after 8.00pm and 9.00pm might well include evening workers and may well be added to by users accessing the Internet from home linking into the site from “dot co” IP providers. What is apparent is that the differences at certain times of the day were considerable. Approximately 25 per cent more views were made to healthy‐eating pages as compared to the share made to depression pages between 10.00am and 12.00pm and between 2.00pm to 5.00pm. The views to depression pages increased by about a third in the period coming up to the lunch break, between 12.00pm and 1.00pm. Views to this page peaked at 1pm then fell back sharply after the lunchtime break. By 3.00pm views to depression pages were about 50 per cent of its lunchtime value. However, use subsequently climbed after 5.00pm. Here there is some evidence that users indeed prefer to search for certain pages, in this case depression, when the office is less crowded: during the lunchtime break and in the period after work. As indicated search disclosure may only be one factor impacting on use here, in fact users in each case might well have different demographic and health interest profiles. However, this is indicative evidence that users may indeed choose their own time to search for sensitive topics, but further research is needed to clarify causality.
The problems of accessing the Internet at work were further highlighted in a qualitative study on Birmingham City Council employees that found the colours of an at work health broadband information system delivered to their work place computers was too strong. In particular, users complained that the distinct colour combination of the service meant that work colleagues knew what was being accessed and this proved to be a “disincentive” to use the service. “Regular” Web sites were described as being less bold and therefore less obvious that you were searching them. Here users were put off accessing a health site if work colleagues knew what they were accessing the service. The work place seems to be an environment were users would like to search with some degree of anonymity and this may well be negotiated or arranged by searching at different times of day or by restricting the topic searched for.
It was not feasible in the current log research programme to separate out home and at work use. However, a questionnaire of NHS Direct Online users asked where they searched from and a comparison between home and work use was made. The percentages of users visiting each section of the Web site, broken down by whether the user searched from work or home, are displayed in Figure 3[2].
It is noticeable that the only health sections where the percentage of home views exceeded the percentage of work views were in the cases of “conditions and treatments” and “listen here”, a facility whereby users can listen to a health information clip. Some users seemed to prefer to search for conditions and treatments in the more private environment of home. This section includes much more sensitive health topics compared to sections such as the “healthy living”, “health in the news” or “healthcare guide” sections. This is further suggestive evidence that people might perhaps want to view some pages in a more anonymous home environment.
Further qualitative research confirmed the home environment as one offering privacy and anonymity. In an online open‐ended type questionnaire study of NHS Direct Online users responses to questions on why the Internet was the chosen medium for procuring health information, suggested that anonymity, alongside the convenience of having information provided in one's own home, were seen as positive Internet attributes. It was interesting that, despite not being prompted (i.e. by multiple‐choice answers), a large majority of the 42 respondents – 26 (62 per cent) – mentioned privacy or anonymity. In some cases it was found, by follow‐up email queries, that “convenience” was interchangeable with “anonymity”. This was also clear from the juxtaposition of the two ideas in the messages of some respondents. One respondent remarked that “It is extremely convenient to use my own PC in the privacy of my home”, while another described the Internet as “A readily available (24*7 at home) and anonymous source of information”. Apart from the linkage of the home with privacy, those who mentioned anything concerned with anonymity or confidentiality did so in somewhat general terms (i.e. such as simply stating that one could look up information privately. One person did confirm that this was for information he felt “unable to ask doctors about”, and another said that the advantages included having no direct contact with health professional.
The importance of searching in a home‐“Internet”‐based environment was further captured by the qualitative study on kiosk mentioned above. One user compared their kiosk search to their Internet search that indicated a preference for the Internet over kiosk in terms of convenience and privacy. In terms of convenience their comment was: “I can sit at home, whenever I like, and surf around” while for privacy their related comment was: “its one thing standing there in front of everyone at a doctor's surgery, and another sitting comfortably at home, in the privacy of your own house, looking up your condition with no‐one peering over your shoulder”.
DiTV
DiTV does perhaps offer users of digital information system the greatest level of privacy and security. Although in the main television sets are situated in public areas of the home, users can choose to a secure time to use the service. DiTV users are searching in their own home and on a medium that that they know and are familiar with.
DiTV transaction logs were examined and a comparison was also made of the use of the pages about depression and healthy living on the Living Health channel over the course of a day, divided by time of day (Figure 4).
The profile of use between depression pages and healthy living pages is quite different. There is an increased incidence of use of depression pages after 6.00pm. The peak use of depression pages is at 9.00pm. Use at this time is about 20 per cent more compared to use of healthy living pages. Note that the pattern of page use on a DiTV system is very different from the pattern of page use of an Internet information service. The recorded swing from the lunchtime peak in use to the fall off in use in the afternoon is much greater on the Internet. There is a much steadier build up in use on DiTV peaking at about midday.
Interview data with the Living Health viewers also indicated, significantly, that time of access was a factor in privacy and usage. Although only a minority of respondents specifically mentioned privacy in the open‐ended interviews, those who did mention it said they generally looked for health information “when the kids are in bed”. One respondent said: “In the morning everyone is out [at work and school] so I can look at anything I want to on the box. I don't really want my 11‐year‐old asking me what ‘period pains’ are”. Four out of 20 users volunteered privacy issues with out the interviewer prompting for reason for use in this context.
A study was conducted specifically to look at the use of sensitive pages between DiTV and other platforms. The purpose was to analyse page usage between platforms to see if there was any variation in use of sensitive pages. It was posited that there would be less usage on the more open and less private kiosk platform compared to either the Web or DiTV. However, there are considerable difficulties in making such comparisons. Kiosk pages for example tended to cover one topic per page whereas Web pages were often grouped so a number of topics were covered by a single “Web” page while DiTV topics tended to be spread over a number of pages. However, a comparison proved possible in the case of HIV topic pages on kiosk and DiTV. And usage did reflect support the supposition. For the kiosks, this sensitive topic just did not receive many page views. Thus the page on HIV received just seven hits out of 223,124 page views and accounted for just 0.0003 per cent of all kiosk content page views. This is a small use and can be compared to the views to an HIV page on the Living Health DiTV information service[3]. Here, 395 hits out of 328,894 page views were recorded to an HIV page and is proportionately substantially higher and accounted for 0.1 per cent of all content page views. Clearly, users were more willing to view an HIV page on a DiTV platform compared to a kiosk. Part of this difference must be accounted for by the anonymity offered by a DiTV service that can be accessed from home compared to a public touch‐screen kiosk.
It was decided also to look at a broad sweep of page views and Table I lists the top 15 pages viewed on each platform. The willingness of DiTV users in the privacy of their own home to view pages of a private nature can be seen more clearly here and the sexual nature of topics searched for (column 1) is plain though curiosity viewing may also play a part here[4]. Furthermore, Web topics are more sensitive or personal than touch‐screen kiosk topics.
The idea that users might be more willing to use DiTV to look for information that they did not want to discuss with other people was further examined in a questionnaire. In a questionnaire study of Communicopia DiTV users respondents were asked about their use of TV for personal and confidential information searching. It asked if respondents looked for information that they didn't want to discuss with their doctor. The responses are given in Figure 1.
A significant proportion (27 per cent) said that they would use the service to look for information that they would not want to discuss with their doctor. An indication that users were attracted to the service for its alleged anonymity and the low search‐disclosure factor. Of course we are talking here about self‐disclosure, in that it says something about the willingness of a person to talk about a private matter with their doctor. However, it also says something about the willingness of users to use online methods, in this case DiTV, to search for private information.
There are significant differences in viewing metrics between the three platforms. Table II provides a comparison of the key time‐based metrics – respectively, the average time that a person takes to view a page/screen, the number of pages in a session and the average length of a search session.
There are significant differences in use between platforms, the log metrics point to a greater use of DiTV. Most notably DiTV users took a third longer to view a screen compared to Kiosk users. Kiosk users are estimated to have the shortest average page view time of approximately 9.1 seconds and Internet users as having the longest at approximately 21.5 seconds. There are a number of reasons explaining these differences and these are further discussed in Nicholas et al. (2003b). The metrics are reported here for information only.
Conclusion
We have amassed significant indicators from across a range of studies to support the notion of search disclosure. Most tellingly we have shown that nearly half of users of a touch‐screen kiosk said that they would not use the service because the system was not sufficiently private Furthermore users interviewed said that privacy and anonymity were important and this was true across all three digital platforms. In addition, patterns of use were observed suggesting that some consumers of more sensitive information such as depression were waiting for periods where search anonymity was guaranteed. This was true both for Internet users and for DiTV users. In particular, the use of an HIV page was substantially higher on DiTV compared to the touch‐screen kiosks. A result that partially reflects the greater anonymity offered by DiTV.
There does seem to be a tendency for users to moderate their online search behaviours as a function of the anonymity afforded at the point of search and information consumption. Under the proposed search disclosure model, touch‐screen kiosks were hypothesised to invoke different patterns of health topic search behaviour from either the Internet or DiTV because kiosks because they offered less search anonymity. There was evidence to support this: users were attracted to both an Internet and DiTV platform because of the privacy and anonymity that they offered. Differences were also found between work and home Internet content search behaviours.
It seems that the willingness of a person to search a digital database for all or certain topics is mediated by the perceived anonymity afforded by the system. This willingness is tied to issues of perceived anonymity and sensitivity about revealing possibly intimate information needs. We have labelled this search disclosure and a concept to help us understand how the user might be discouraged or encouraged to use a particular information system in a particular location for a particular topic.
There has been a revolution in who uses and searches for information brought about by new consumer driven digital information systems. Issues of privacy, convenience, ease of use and digital visibility are driving this revolution. In this paper we have sought to explore one characteristic – “search disclosure”. We believe that this is exceedingly important in the context of understanding existing consumer search behaviour and also in laying out a methodology for analysing online services that may offer a more personalised online service, such as electronic patient record keeping.
What we have presented here is some early thinking on an information‐seeking concept that we call social disclosure. Clearly, we need now to conduct separate “controlled” studies comparing use in different situations. For instance, it may be possible to conduct experiments in controlled environments in which participants search online content on a touch‐screen kiosk with other present or absent. Observational research in the field could be used to monitor closely different online search behaviours on kiosks at times when others are present or not. However, it is not clear that we can “control” for all the variables that impact on kiosk or digital system use. It has to be said that all log‐derived data is questionable, as we cannot separate out the different effects on use from search disclosure. Search disclosure will always be one of many effects but probably significant in fields like health and in the cases of certain “questionable” consumer interests (pornography for instance). The picture provided by the logs is inevitably subject to confounding and ambiguity. For instance there is a type of information seeking which is more akin to browsing, what people will look for in an idle moment. A kind of curiosity search. Curiosity being a reason to look at sensitive “sex” type information, for instance. However, a search of the literature was conducted on the likely impact of “curiosity” viewing but no logical model to support how “curiosity” might interplay with search disclosure could be identified. There is also the remaining issue of PC versus DiTV use. We need to compare search behaviours for DiTV and for home PC Internet users. At the moment, data on PC use aggregates over potential home and work use. However, we do have particularly “good” evidence supporting the argument for search disclosure from both questionnaire data and qualitative research. We are not basing the analysis just on one method. Each method alone is not sufficient in arguing the case for search disclosure but together they paint a pretty convincing picture. You have to decide if this picture is good enough.
Notes
A total of 27 days; however, the system was down for one day (11 September), so the period covered data for 26 days only
All difference significant at the 5 per cent significance level.
This comparison is between use of an HIV‐labelled page on both platforms. Living Health also had additional HIV‐related pages including HIVAIDS, HIVAsia, HIVVaccine and hepatitusHIV; however, these were not included in this analysis.
Curiosity viewing was thought to play a part in the record hits to sex information pages on the DiTV platform. However, the information was comprehensive and this may just as well reflect a real demand for this information.
References
Coleman, L.H., Paternite, C.E. and Sherman, R.C. (1999), “A re‐examination of deindividuation in synchronous computer‐mediated communication”, Computers in Human Behaviour, Vol. 15, pp. 51‐65.
Huntington, P., Nicholas, D., Williams, P. and Gunter, B. (2002), “Comparing two digital consumer health television services using transactional log analysis”, Informatics in Primary Care, Vol. 10, pp. 147‐59.
Joinson, A. (1999), “Social desirability, anonymity and Internet‐based questionnaires”, Behaviour Research Methods, Instruments and Computers, Vol. 31 No. 3, pp. 433‐8.
Joinson, A.N. (1998), “Causes and implications of disinhibited behaviour on the Internet”, in Gackenbach, J. (Ed.), Psychology and the Internet, Academic Press, San Diego, CA, pp. 43‐60.
Joinson, A.N. and Banyard, P. (2002), “Psychological aspects of information seeking on the Internet”, ASLIB Proceedings: New Information Perspectives, Vol. 54 No. 2, pp. 95‐102.
Julien, H. (1999), “Where to from here? Results of an empirical study and user‐centred implications for system design”, in Wilson, T.D. and Allen, D.A. (Eds), Exploring the Contexts of Information Behaviour, Taylor Graham, London, pp. 586‐96.
Leydon, G.M., Boulton, M., Moynihan, C., Jones, A., Mossman, J., Boudioni, M. and McPherson, K. (2000), “Cancer patients’ information needs and information seeking behaviour: in depth interview study”, British Medical Journal, Vol. 320, pp. 909‐13.
McIntosh, J. (1977), Communication and Awareness in a Cancer Ward, Croon Helm, London.
Nahl, D. and Tenopir, C. (1996), “Affective and cognitive searching behaviour of novice end‐users of a full text database”, Journal of the American Society for Information Science, Vol. 47, pp. 276‐86.
Nicholas, D., Huntington, P. and Williams, P. (2001), “Establishing metrics for the evaluation of touch screen kiosks”, Journal of Information Science, Vol. 27 No. 2, pp. 61‐72.
Nicholas, D., Huntington, P. and Williams, P. (2002a), “The impact of location on the use of digital information systems: case study health information kiosks”, Journal of Documentation, Vol. 58 No. 3, pp. 284‐301.
Nicholas, D., Huntington, P., Williams, P. and Gunter, B. (2002b), “First steps towards providing the nation with health care advice and information via their television sets: an evaluation of pilot projects exploring the health applications of digital interactive television”, report presented to the Department of Health, London, August.
Nicholas, D., Huntington, P. and Williams, P. (2003a), “The characteristics of users and non‐users of a kiosk information system”, working paper, City University, London.
Nicholas, D., Huntington, P. and Williams, P. (2003b), Working paper, to be included in final report of The Web, the Kiosk, Digital TV and the Changing and Evolving Face of Consumer Health Information Provision: A National Impact Study, Department of Health, London, 2000‐2003.
Williams, P., Nicholas, D. and Huntington, P. (2003), “Non‐use of health information kiosks examined in an information needs context”, Health Information and Libraries Journal, May.
Wilson, T. (1995), “‘In the beginning was the Word …’: social and economic factors in scholarly electronic communication”, Department of Information Studies, University of Sheffield, Sheffield, available at: www.shef.ac.uk/∼is/wilson/publications/elvira.html (accessed 29 July 2003).