Abstract
Purpose
The purpose of this paper is to report an empirical investigation into conceptions of the “invisible web”.
Design/methodology/approach
This was an exploratory qualitative study based on in‐depth semi‐structured interviews with 15 members of academic staff from three biology‐related departments at the University of Sheffield. Concepts emerged from an inductive analysis of the interview data to form a tentative model.
Findings
A distinction is drawn between technical objective conceptions of the “invisible web” that commonly appear in the literature, and a cognitive subjective conception based on searchers' perceptions of search failure, and a tentative model of “cognitive invisibility” is presented. The relationship between objective and subjective conceptions, and implications for training, are discussed.
Research limitations/implications
The research was qualitative and exploratory, designed to elicit sensitising concepts and to “map the territory”. It thus aims to provide a tentative model that could form the basis for more systematic study. Such research could investigate the validity of the categories in different and/or larger samples, seek further to illuminate, challenge, extend or refute the model, and address issues of generalisability.
Practical implications
The paper presents a conceptual model that is intended to be a useful reference point for researchers wishing to investigate user‐based aspects of search failure and the invisible web. It may also be useful to trainers and those interested in developing information literacy, in that it differentiates technical objective and cognitive subjective conceptions of “invisibility, and discusses the implications for helping searchers develop more effective searching capabilities.
Originality/value
The paper offers an alternative cognitive subjective view of “web invisibility” to that more commonly presented in the literature. It contributes to a still small body of empirical research into user‐based aspects of the invisible web.
Keywords
Citation
Ford, N. and Mansourian, Y. (2006), "The invisible web: an empirical study of “cognitive invisibility”", Journal of Documentation, Vol. 62 No. 5, pp. 584-596. https://doi.org/10.1108/00220410610688732
Publisher
:Emerald Group Publishing Limited
Copyright © 2006, Emerald Group Publishing Limited
Background
“Invisible” web resources consist of material that general‐purpose search engines either cannot, or are not intended to index. Minkle (2002) called such invisible resources the “buried treasure of the web”. He noted that this portion of the web is not totally invisible – but rather, invisible only to users of conventional web‐based search engines. As Devine and Egger‐Sider (2004) state:
Sherman and Price (2001) comment that the invisible web is huge, and in all likelihood is growing more rapidly than the visible web. Pedley (2002) also notes the size and quality of information that “resides” in this area of the web and which may thus be missed. Sherman and Price (2001, 2003) group the invisible web into four main categories according to possible reasons for invisibility: the opaque, the private, the proprietary, and the truly invisible web. These are shown in Figure 1.… the invisible web comprises all the information sources available on the web that are overlooked by conventional search engines.
The opaque web
The opaque web comprises resources that could be included in search engine indexes but are not. There may be several reasons for this, including depth and frequency of crawling. Some providers adopt a policy of “good enough” rather than deep crawling, the result of which is that information is missed beneath certain hierarchical website levels. Search engines also differ in the way they display search results, usually limiting them in some way. Also, there is inevitably some delay between information being added or changed and its appearance in a search engine's index. This is particularly problematic with current, real‐time or constantly changing information – for example, company stock price, news and weather forecasts. Pedley (2001) termed this area the vanishing web. Another part of the opaque web in Sherman and Price's categorization relates to search engine policies for displaying search results. For example the query “biological research in the UK” at the time of searching in Google produced 8,530,000 results, of which 663 were displayed with an option on the final page to display a further 241. Studies such as that by Jansen et al. (2000) indicate that the majority of search engine queries are short, advanced search facilities including Boolean operators (which might reduce potentially the very large hit lists that might result from such short queries) being seldom used.
The private web
The private web refers to resources that technically can be indexed but which are deliberately excluded from search engine crawlers. Password protection, the robots.txt file and the noindex meta tag are three common methods of preventing access by crawlers. This part differs from other parts of the invisible web in that these resources are not intended by their owners to be visible to general searchers.
The proprietory web
The proprietary web refers to resources that require registration for access. Some of these sites are fee‐based, but many are free of charge. However, in both cases, search engine crawlers are not able to satisfy the registration requirements and cannot access these resources.
The truly invisible web
As Hsieh‐Yee (2001) notes, most search engines are mainly geared to indexing static HTML pages. However, such pages represent just a small fraction of the information available on the web. The part of the web which is not in static format and cannot be indexed by general purpose search engines has been termed the truly invisible web. This includes resources that are in formats that can not be indexed by certain search engines, dynamically generated web pages, and information stored in relational databases. Such dynamically generated information has been called the deep web (Bergman, 2001). A number of search engine crawlers cannot index resources such as postscript files, flash, audio, and streaming video. The volume of such resources is substantial, and this can represent a serious limitation of search engines. Increasingly, however, mainstream search engines – including AltaVista, Google amd Yahoo – are able to index a variety of other formats.
Other factors
Another type of invisibility, not mentioned by Sherman and Price, relates to the fact that the majority of web users appear to view a very limited number of result pages. Spink (2003), for example, estimates that searchers view on average only some 2.35 pages; whilst Jansen and Pooch (2001) report that web searchers typically view no more than ten results.
What have been termed disconnected URLs (Sherman and Price, 2003) represent another aspect of the invisible web. These include web pages that have not been submitted to a search engine and have no link pointing to them. Broken links are another problem, and can result from web pages being deleted or moved. There is usually a delay between moving a page and the updating of search engine indexes. Therefore, even if pages still exist elsewhere on the web they remain in the invisible realm. Block (2004) terms this area of the web the dead web. Broder et al. (2000) estimated that disconnected URLs made up about 20 per cent of the potentially indexable web.
An empirical study
Hsieh‐Yee (2001, p. 181) noted that:
… web search behaviour is a fertile ground for research … Researchers can pursue nearly any aspect of searching they wish. But … they need to place users at the centre of investigation, understand users' needs and find out how the current web environment supports or hinders their information seeking.
However, although the term “the invisible web” appears frequently in the literature (e.g. Sherman and Price, 2001, 2003; Pedley, 2001, 2002; Minkle, 2002; Clyde, 2002; Crowther et al., 2004), little if any non‐technical user‐based empirical research has been published. Rather, the invisible web has been considered from a predominantly technical perspective in terms of the reasons for different types of invisibility. The research reported here, however, is based on empirical data relating to information seekers' perceptions of their search experiences.
Research questions
The study reported here was an exploratory investigation into academic users' conceptions of the “invisible web, and into their normal experiences of web searching, with particular reference to perceived search failure. The research questions addressed were:
- •
How do a particular group of users conceive the “invisible web”?
- •
What are users' experiences of search failure on the web?
- •
To what extent do such conceptions and experiences map onto the technical models of the invisible web proposed in the literature (such as that shown in Figure 1)?
Methodology
A qualitative approach was adopted for the study, based on the inductive analysis of a series of 15 in‐depth semi‐structured interviews. Exploratory inductive research has been conducted in relation to other aspects of web searching (e.g. Wang et al., 2000; Ford et al., 2002, 2003, 2005a, b), such studies aiming to identify sensitising concepts and to suggest potential lines of further systematic enquiry. This study also sought to identify sensitising concepts – to establish a tentative map of ways in which users conceive of and experience aspects of the invisible web.
Sampling
The sample consisted of 15 academic staff from the departments of Animal and Plant Sciences, Biomedical Science and Molecular Biology and Biotechnology at the University of Sheffield. There were three main reasons for the selection of this particular community for the study:
- 1.
A literature review suggested that the information seeking behaviour of biologists – in particular web‐based information seeking – has received little attention to date.
- 2.
There is a large community of biologists in Sheffield, the three selected departments comprising one of the largest in the UK.
- 3.
The researcher's subject background is in agricultural science and he is familiar with the relevant subject areas.
All members of academic staff of the three departments were contacted individually by letter, 15 agreed to participate, and all were interviewed. This sample represented 12.5 per cent of the total 120 members of academic staff.
Data collection
Interviews were conducted between 5 November 2004 and 12 January 2005. They were held in the offices of the participants and recorded digitally. The average duration of the interviews was 40 minutes, ranging from a maximum of one hour and five minutes to a minimum of 28 minutes. Critical incident technique was adopted as a means to facilitate data collection. This worked well and without exception interviewees were able to remember a recent search in detail. All interviews were fully transcribed, each interviewee being sent her/his transcript to enable confirmation of its accuracy, and to provide an opportunity for changes and comment.
The main interview schedule is listed below (Q1‐Q20): Q1. What is your general feeling about searching the web? How do you find it in general? Q2. How long have you been searching the web? Q3. How often do you search the web? Q4. What kind of information do you usually look for on the web? Q5. How satisfactory are search results for you in general? Q6. Can you remember and describe one of your recent experiences in which you were successful in web searching? Q7. Why do you think in this specific case you have been successful? Q8. Do you always manage to find what you want on the web or do you ever not find what you have been looking for? Q9. Can you remember and describe one of your recent experiences in which you failed to find what you have been looking for? Q10. Why do you think this specific case was a failure? Q11. How do you feel about your information seeking failure on the web? Q12. How do you differentiate between failure and success in searching the web? Q13. Do you ever get the feeling that there should be more relevant information about your search topic on the web but you are not getting to it? Q14. Do you think that you always find everything that exists on your search topic on the web? I mean, in your opinion how likely is it that you have missed something about your search topic? If so, why couldn't you find it? Q15. How much does it matter to you if you know you have missed something while searching the web? How do you feel about it? … Do you always feel like that? Q16. When you fail to find what you want on the web, what do you do about it? I mean, how do you seek different solutions to cope with information seeking failure on the web? Q17. How did you develop these coping strategies? Before ending the interview, I need to clarify two more things:Q18. What does the web mean to you? I mean, what do you mean by the web? Q19. Could you describe what “The Invisible web” might mean to you? Q20. Is there anything else that you would like to add? During the interviews, the researcher also made use of active listening and paraphrasing techniques, coupled with more detailed probe questions where necessary. Since the nature of such probes were context‐dependent relating to particular interviews, they are not included above.
Data analysis
The interview transcripts were analysed inductively in accordance with the procedures recommended by Glaser and Strauss (1967). The intention was to avoid any imposition of preconceived ideas on the data, and to ensure that any emergent interpretations were well grounded in the data. The recommendation of Strauss and Corbin (1998) first to employ word‐by‐word “micro analysis” was rejected on the grounds that the level of granularity of the resultant analysis would represent – to quote Glaser's (1992) criticism – “over‐conceptualisation”. Rather, key points were identified, coded and analysed to form higher level concepts. Glaser and Strauss' constant comparison approach was used until it was felt that saturation had been reached. Higher order categories emerged from the conceptual analysis, and relationships between these were explored, resulting in the tentative model presented in the next section.
Results
Characteristics of the sample
The sample consisted of three professors and one lecturer from the Department of Animal and Pant Sciences; one professor and four senior lecturers from Biomedical Science; and three professors, one reader, one research fellow and one lecturer from Molecular Biology and Biotechniology. The average age of the interviewees was 47.9, ranging from 62 to 31.
Fourteen of the interviewees had engaged in web searching more or less since the web had become accessible at the university. The one exception started searching three years ago. All typically conducted multiple searches daily.
Responses to the question “what kind of information do you usually look for on the web?” revealed a variety of topics, including work‐related and personal. The following quotations give a favour of the wide range of academic‐related searching activity:
To see what research people do… looking for papers for information and scientific subjects.
Journal articles, technical resources for lab techniques, and buy chemicals things to buy for our lab, DNA resources and then pictures for putting into presentations and then background research information.
Information to support my teaching, but in terms of the scientific content, but also perhaps issues to do with learning and teaching – what is the best way of getting this information across to the students rather than what is the scientific information.
When I am writing a paper then I would maybe be looking for relevant references … When I am writing lectures, and PowerPoint presentations these days, then I am looking for good pictures to go with the data.
The term “invisible web” was new to all but one of the interviewees:
What the invisible web might mean? Nothing, it doesn't mean anything to me. I don't know. I have never heard of the term it doesn't mean anything to me … I don't know unless this term is used by bioinformatics people for pages that one can't access because you can't find the correct search word.
Hmm, I don't know. It could it be information that is there [on the web] that people don't use – perhaps things that are just dumped there for all sorts of purposes, but nobody ever accesses, such as minutes of committee meetings and that sort of thing. I have to say I am struggling with it and I don't know what it means. What is it?
“Cognitive invisibility”
Although the term “invisible web” was new to all but one of the interviewees, when asked to recall a “successful” and a “failed” search, all could do so without hesitation. Interviewees were asked to describe a successful and a failed search, and a number of recurring themes emerged from the interview transcripts, relating to the extent to which the interviewees felt that information relevant to their queries was “out there” – i.e. was available to be retrieved if its URL could be found. These themes are categorised and presented as a tentative model shown in Figure 2. The model is first presented, followed by an explanation of the categories along with short quotations illustrative of the interview transcripts on which the categories are based.
Figure 2 shows a categorisation of the different themes emerging from the interviews in the form of a “matrix of cognitive invisibility, beginning in the top left corner with total visibility and going on to include gradations of invisibility. It is important to note that the notion of “invisibility” depicted here is different from that discussed in the section 1 of this paper. The matrix shown in Figure 2 represents the subjective views of the interviewees. It is important to note that the interviewees' perceptions of whether information they were searching for was indexed and available for retrieval do not necessarily (although they may) correspond to the reality of whether it actually was so indexed and available (and could possibly have been found by other – for example more expert – searchers). Invisibility is as perceived by the interviewees and is thus a cognitive artefact. This introduces a cognitive subjective, as opposed to a technical objective definition of “invisibility”.
Thus, “invisibility” as defined here relates to information sources that the interviewees failed to find for whatever reason including, for example, the use of inappropriate search terms. Thus, what may have been invisible to one searcher may not necessarily have been invisible to another. “Search failure” is similarly a cognitive artefact, representing searchers' perceptions that a search has been unsuccessful (however they define this). The “levels of uncertainty” (referred to in the matrix), that information was indexed and potentially available to be retrieved via a search engine, are inferred from the interview data, and explained in the subsections below.
The arrows indicate that, although discrete areas (A1‐B3) are shown in the matrix, the differences are essentially of degree – whether of uncertainty that there is relevant information indexed and available for retrieval, or of relevance.
Bright zone
To set the notion of failure, and degrees and types of failure in context, it is worth noting that all interviewees could recall search experiences that lacked any perceived level of failure – and indeed difficulty – entailing the straightforward retrieval of relevant information via a search engine without problems:
That [search] was very straightforward and I had a very efficient and rapid answer …
I wanted to find out more about [my friend's new] department and so I did a Google search in that case and typed in the name of the university department and went straight to it … I typed in the words in inverted commas and I … went straight to the department and it was very successful.
In zone A1, the searcher searches and finds relevant information using a search engine without problems. This zone is terms bright since the search is fully illuminated with no hindrance.
Refracted zone
In zones A2 and A3, the searcher finds relevant information by means other than a search engine via bookmarked websites, and/or URLs given by friends and colleagues. Although not representing search failure, these experiences are considered relevant here insofar as they represent the bypassing of search engines, and search engine use is central to the concept of both technical and cognitive conceptions of “invisibility” that form the focus of the present study. This category thus represents a gradation in the move from total visibility to total invisibility shown in the emerging model presented below.
The question of whether the information retrieved could have been retrieved via a search engine (indeed, bookmarked sites could have been so retrieved originally) may simply not be known by – or indeed relevant to – the user, and thus not referred to in the interviews (category A2):
Well, there is one example, yes … It was a colleague who told me that it is there somewhere … I would e‐mail my colleague and say “Can you give the URL?”
I was looking for … full spectrum lights … I couldn't find it and eventually a friend found it for me through another source – and they had a website. … She didn't find it on the Internet; she actually found it fortuitously in a leaflet she has got. So, she found it that way.
Occasionally, however, the user indicated surprise at finding information, presumably indicated an initial feeling that the information would not be indexed (category A3):
I didn't think I'll be able to find the information, and … I was rather heartened and surprised that I did find it.
Veiled zone
Inevitably, however, there were perceived searching failures. In some cases, the interviewee did not only fail to retrieve the information hoped for, but also felt confident that it was there to be retrieved (category B1). The user searches expecting to find information via a search engine but realises that s/he has failed to locate it. This zone is termed veiled since information is felt to be there but somehow concealed from view:
I'm sure the information is out there, and I'm sure this is the case for most things – that information is there you just have to spend a bit longer looking around, or maybe using more directed search engines or something like that.
I know the information must be somewhere and I know if I did it properly I will be able to find it but I haven't yet found it.
Dark zone
Sometimes, however, interviewees were very unsure whether or not the information was “out there” and retrievable with greater effort or skill on their part (category B2):
But the problem is I don't know if it is there and it is the search engine which fails, or if I am putting the wrong information or the information isn't there.
It's hard to know, you never know, whether you can't find it because it is not there or if you can't find it because your search method is inappropriate … You don't know when to stop, when to give up, because you don't know if you are wasting your time and there is literally no information out there online or whether you just not searched appropriately …
Towards the other extreme, whilst not actually knowing the reason for certain failed searches, interviewees reported occasions when they felt confident that information was not “out there” indexed and ready to be retrieved (category B3):
Failure is probably mostly generated by the fact that the information just isn't there. Then so what? We just carry on, so it is not a big deal.
Yes it does fail on occasion but that is probably because the information isn't there rather than being hidden … When I am looking for things if I don't get anything then I am pretty certain that it is not there.
Thus in the dark zone, the user fails to find relevant information but either simply does not know whether or not it was there to be found (B2), or feels that in all likelihood the information was not there to be found (B3). The latter situation could pertain to a search specifically aiming to confirm that no relevant information is “out there” – for example, on a new research topic. Alternatively, the conclusion that there is no information to be found may be drawn after the search. Presumably in the first case, there was at least a small suspicion that they may have been information, otherwise the search would be unlikely to have been conducted in the first place.
Discussion
The matrix of cognitive invisibility is arguably a useful complement to more objective technical conceptions of invisibility in that the latter may be problematic for the researcher interested in user issues. This derives from the fact that what constitutes the objective invisible web is a moving target since it is subject to constant change. What is invisible today may not be so tomorrow. As one interviewee put it:
You look at it this morning and this web looks like this and then look at it tonight the web has grown, the hub's grown and some more things and more connections but equally some have gone.
The capability of search engines is also constantly being improved, affecting objective visibility. For example, file formats that were not crawled some years ago are now visible to many general search engines.
Also, each user's search skills are likely to change with experience over time. For this reason, it is important to distinguish, and to be clear about the relationship between, what we have termed here the “objective technical” invisible web and the “cognitive subjective” invisible web. A simple example illustrates the distinction. A user searches for but fails to find information on a topic. In reality, there is no information “out there, but the user feels confident that there is, and that s/he has failed to find it (category B3 in Figure 2). This example does not relate to the objective invisible web, since in reality there is no web‐based information “out there” rendered invisible by, for example, search engine indexing practice or policy. However, it does relate to the cognitive invisible web, since the searcher clearly perceives that s/he has to some extent failed to find what was there to be retrieved.
This is arguably important in that there is an interplay between searchers' perceptions of search failure, and the objective reality of search failure. This interplay may affect searchers' confidence, and has implications for training. Perceptions may not relate to reality, and vice versa. A user's perceptions may help to determine his or her confidence, and her or his motivation and perceived need for self improvement thus impacting on training. The reality may affect the nature and extent of training that may actually be required to render the user a more effective searcher in that training should reference the “reality” of a searcher's skill as well as her/his perceptions. Therefore both objective and subjective conceptions of the invisible web, and interaction between them, may be useful. It is likely also that the perceived importance of a failure is a factor in such processes – failure to find information in relation to a relatively trivial enquiry having less impact than failure in a search of significant importance to the user.
Drawing such a distinction between the objective and perceived web enables some mapping of a searcher's skills onto “reality”. This mapping may be useful in terms of helping the searcher to develop appropriate expectations, confidence, and improved search skills. It is acknowledged that an objective assessment of any individual's “actual” invisible web may be unknowable with any accuracy, the nearest approximation being the assessment of an experienced web searcher. Nevertheless, such an approximation may still be useful.
For example, were a searcher to perceive that search failure was not due to any lack of skill on his/her part, confidence may be beneficially affected – or, at least, detrimental effects may be avoided. However, if we are to help people develop skills that are likely to result in more effective searching, we need to map such perceptions onto as realistic a view as possible of these skills. Thus more objective assessments may be useful in the process of helping people develop more effective search skills. Such a mapping may be instrumental in helping to ameliorate both excessively negative misperceptions on the part of a searcher in relation to her/his abilities (which may lead to a lack of confidence and motivation), and excessively positive misperceptions (by suggesting the need for appropriate self‐critical reflection).
Conclusions
The categories shown in the Figure 2 matrix were derived inductively from the interviews, and account for all the relevant interview data, in the sense that no contradicting or competing data has been ignored. Whilst it is argued here that the matrix validly represents the range of search experiences reported by the academic staff interviewed, the research did not seek to investigate the frequency with which any of the six types of search experience depicted in the matrix occurred, or are likely to occur in any wider population. Rather, it aimed to chart the possibilities, or identify what Olaisen (1991) has termed “sensitising concepts”. The function of this type of research is to “map the territory” and provide at least a tentative scheme that could form the basis for more systematic study.
Such subsequent study could, for example, investigate the validity of the categories in different and/or larger samples, and seek further to illuminate, challenge, extend or refute the categories. If the model were further developed, the frequency of occurrence of different types of experience in different populations and issues of generalisability could also be investigated using statistical methods.
The explicit notion of “the invisible web” meant little to all but one of the interviewees. Nor did any of the aspects of cognitive invisibility derived from the interview data map in any detail onto any of the aspects of technical invisibility shown in figure 1 – except for a general feeling that information may or may not be indexed and thus retrievable by search engines. However, despite this lack of mapping, the notion of “invisibility” is nevertheless arguably useful as a metaphor for describing differences in users' search failure perceptions, as in the matrix presented above.
Whilst as a metaphoric term, “invisibility” bridges the objective technical search engine realities that may bring about search failure, and subjective perceptions of such failure by users, like any analogy it exhibits differences as well as similarities as used in relation to these two areas of study. However, potential confusion is arguably outweighed by the power of such a common denominator to emphasise the close links and inter‐dependencies of technical and cognitive aspects of search failure, and the importance of mapping such links and inter‐dependencies to support education and training to help people become more effective in their information seeking.
About the authors
Nigel Ford is a Professor in the Department of Information Studies, University of Sheffield, Sheffield, UK. He is the corresponding author and can be contacted at: [email protected]
Yazdan Mansourian is a PhD Researcher in the Department of Information Studies, University of Sheffield, Sheffield, UK.
References
Bergman, M.K. (2001), “The deep web: surfacing hidden value”, The Journal of Electronic Publishing, Vol. 7 No. 1, available at: www.press.umich.edu/jep/07‐01/bergman.html (accessed October 6, 2005).
Block, M. (2004), “The invisible web”, a presentation for LACONI‐RASS at Batavia Public Library, available at: http://marylaine.com/hidden.html (accessed October 6, 2005).
Broder, A., Kumar, R., Maghoul, F., Raghavan, P., Rajagopalan, S., Stata, R., Tomkins, A. and Wiener, J. (2000), “Graph structure in the web”, presented at the 9th International World‐Wide Web Conference, available at: www.almaden.ibm.com/cs/k53/www9.final/ (accessed October 6, 2005).
Clyde, A. (2002), “The invisible web”, Teacher Librarian, Vol. 29, pp. 47‐9.
Crowther, P., Bechhofer, S. and Horan, B. (2004), Shedding Light on the Hidden Web, Reprint Series CSPP 28, University of Manchester, Department of Computer Science, Manchester, UK, available at: www.cs.man.ac.uk/cspreprints/PrePrints/cspp28.pdf (accessed 6 October 2005).
Devine, J. and Egger‐Sider, F. (2004), “Beyond Google: the invisible web in the academic library”, Journal of Academic Librarianship, Vol. 30 No. 4, pp. 265‐9.
Ford, N., Miller, D. and Moss, N. (2002), “Web search strategies and retrieval effectiveness: an empirical study”, Journal of Documentation, Vol. 58 No. 1, pp. 30‐48.
Ford, N., Miller, D. and Moss, N. (2003), “Web search strategies and approaches to studying”, Journal of the American Society for Information Science and Technology, Vol. 54 No. 6, pp. 473‐89.
Ford, N., Miller, D. and Moss, N. (2005a), “web search strategies and human individual differences: cognitive and demographic factors, internet attitudes, and approaches”, Journal of the American Society for Information Science and Technology, Vol. 56 No. 7, pp. 741‐56.
Ford, N., Miller, D. and Moss, N. (2005b), “Web search strategies and human individual differences: a combined analysis”, Journal of the American Society for Information Science and Technology, Vol. 56 No. 7, pp. 757‐64.
Glaser, B.G. (1992), Emergence versus Forcing: Basics of Grounded Theory, Sociology Press, Mill Valley, CA.
Glaser, B.G. and Strauss, A.L. (1967), The Discovery of the Grounded Theory: Strategies for Qualitative Research, Aldine de Gruyter, New York, NY.
Hsieh‐Yee, I. (2001), “Research on web search behaviour”, Library & Information Science Research, Vol. 23 No. 2, pp. 167‐85.
Jansen, B.J. and Pooch, U. (2001), “A review of web searching studies and a framework for future research”, Journal of the American Society for Information Science and Technology, Vol. 52 No. 3, pp. 235‐46.
Jansen, B.J., Spink, A. and Saracevic, T. (2000), “Real life, real users and real needs: a study and analysis of user queries on the web”, Information Processing and Management, Vol. 36 No. 2, pp. 207‐27.
Minkle, W. (2002), “The invisible web”, School Library Journal, Vol. 49 No. 12, p. 29.
Olaisen, J. (1991), “Pluralism or positivistic trivialism? Important trends in contemporary philosophy of science”, in Nissen, H.E., Klein, H.K. and Hirschheim, R. (Eds), Information Systems Research: Contemporary Approaches and Emergent Traditions, Elsevier, Amsterdam, pp. 235‐65.
Pedley, P. (2001), The Invisible Web: Searching the Hidden Parts of the Internet, Aslib, London.
Pedley, P. (2002), “Why you can't afford to ignore the invisible web”, Business Information Review, Vol. 19 No. 1, pp. 23‐31.
Sherman, C. and Price, G. (2001), The Invisible web: Uncovering Information Sources Search Engines Can't See, CyberAge Books, Medford, NJ.
Sherman, C. and Price, G. (2003), “The invisible web: uncovering sources search engines can't see”, Library Trends, Vol. 52 No. 2, pp. 282‐98.
Spink, A. (2003), “Web search: emerging patterns”, Library Trends, Vol. 52 No. 4, pp. 299‐306.
Strauss, A.L. and Corbin, J. (1998), Basics of Qualitative Research Techniques and Procedures for Developing Grounded Theory, Sage, Thousand Oaks, CA.
Wang, P., Hawk, W.B. and Tenopir, C. (2000), “Users' interaction with world wide web resources: an exploratory study using a holistic approach”, Information Processing and Management, Vol. 36 No. 2, pp. 229‐51.