Introduction

1.1 Variations in citations rates across disciplines

1.2 Academic careers and the accumulation of citations

1.3 Different career trajectories and publications profiles

Summary

_________________________________________________

Introduction:

Understanding why citations patterns are the way they are for any individual academic or researcher, and how they might be improved upon, is not a simple thing to do. Thinking about these issues demands a good deal of appropriate context. Some academics may avoid dipping their toes into the water at all out of fear that their work is not being cited as much as they would like, whereas others are more keen to better understand their citation record. The purpose of this chapter then is to give readers an appropriate context, within which it will be easier to make sensible judgements about citations. How many cites anyone can expect to get depends on several key factors – especially which country they work in, how old they are (or how far out they are from their PhD), other factors like gender or career interruptions, the type of specialised academic role that their career fits into, and the distinctive features of their discipline and sub-discipline.

Another concern with citations is that people think academics may get cited as easily for making a famous mistake as for getting something right. In principle this is possible – but in practice we know that academics do not usually cite mistakes or work that they believe is plain wrong. Our team looked at 10,400 citations in social science papers and found that explicitly negative commentaries accompanying a citation occurred in only 10 out of these cases. If a paper is wrong (or thought to be wrong), it is simply not cited. If an academic cites research from work with which she disagrees, however, we believe this achieves just as much impact as being cited because the academic fully agrees with the author.

It is tricky for an individual academic to make sense of their citation record but it is even more difficult for a whole department or research lab to understand how they are collectively performing or to make sensible judgements about what more they could or should do to create greater impact. Academics who are asked to sit in judgement on colleagues – whether because of interviewing applicants for new staff roles, appraisal or promotion systems, mentoring, or departmental administration tasks – should therefore take special care to appreciate the complexities described in this chapter. There is no realistic single archetype of how an academic career should develop but rather a number of different trajectories. This diversity both reflects the diverse talents and capacities of academics and researchers themselves, and it responds to the complex, interlocking needs of disciplines, departments and research labs for many different types of contributions.

We begin by considering the different rates at which publications are cited across disciplines- namely the STEM subjects (science, technology, engineering, and mathematics) the social sciences and humanities. Within these gross differences in citation rates, we turn secondly to look at the overall influence of age and experience in shaping the accumulation of citations. Third, many different factors at work across an academics’ lifetime – such as their choices of what to do, their experience or their success in getting to a research intensive university – can be summarised by considering a number of somewhat stylised career trajectories. In the third section of the chapter we consider how these narratively organised influences shape characteristic publications profiles and citation rates.

1.1 Variations in citations rates across disciplines

The average article in the social sciences and humanities is cited less than once a year. Anne-Wil Harzing (2010)

For many years (from 2004 to 2009) the leading UK specialist magazine for the university sector, the Times Higher Education or THE, published league tables of world universities that purported to show their academic quality derived from their gross citation counts. In fact, the rankings by THE principally showed how large their medical faculties and physical science faculties were relative to other parts of the universities, as universities with big medical schools and lots of staff in the physical sciences did very well. And universities without them did relatively worse. Yet the power of focusing on what is easily or immediately quantifiable was such that it took many years for THE to admit that their approach was deficient, fire their citations metrics analysts and recruit a new team.

Figure 1.1 shows the roots of this problem by looking at the total number of citations to journal articles in a given year divided by the number of journal articles produced in the same year, as recorded in ISI Web of Science (ISI), the online academic citations database published by Thomson Reuters. The cite rate in medicine is greater than the cite rate in the social sciences by a factor of 8 to 3, and greater than that in law and the humanities by a factor of 8 to 1. Physical sciences papers in the ISI are also cited twice as often as those from the social sciences, and four times as often as those in law and the humanities.

Figure 1.1: Differences in the average aggregate citation rates between major groups of disciplines, (that is, total citations divided by number of publications


Source:
Centre for Science and Technology Studies (2007)

There are many possible reasons for this patterning. In medicine all published papers are written to a word limit of 3,000 words, whereas the norm in the social sciences is for main papers to be around 6,000 to 9,000 words long. Medical sciences have also developed a strong and rigorous culture of ‘systematic review’ which requires that all relevant studies be cited initially, but that only those that pass certain criteria for methods and merit need be analysed closely.  This very structured and well-defined approach to reviewing literature is mirrored (perhaps in a less rigorous way) in the physical sciences. But a culture of systematic review or comprehensive referencing is far from being established in most social science disciplines – for instance, in theoretical economics and public choice only methodologically similar work is cited, and authors often make a cult of minimal referencing. Systematic review, or a stress on comprehensive referencing, is entirely absent in the humanities.

 

The differences in citation patterns between the medical/physical sciences and the social sciences and humanities can also be explained by the development of a ‘normal science’ culture in the former – whereas in the social sciences there are still fundamentally opposed theoretical streams across most of the component disciplines. In the social sciences citations can become a way of taking sides on what constitutes a valid argument. All of these features are even more strongly marked in the humanities, where referencing is often a matter of personal choice.

While discussing citation patterns it is worth briefly mentioning several technical facts about ISI coverage which will discussed in more detail in Chapter 2. First, its roster of journal articles is much more comprehensive for the medical and physical sciences, the areas where the database first developed. Second, the ISI does not include books (which are an important element of professional communication in the social sciences and the humanities), but does include book reviews (very much more important in the social sciences and humanities, but of course almost never cited by anyone else and hence tend to depress the average citation scores of these disciplines). Third, we know that the self-citation rates (where academics cite their own work) vary dramatically across disciplines – for instance, being twice as high in engineering as they are in political science (see Chapter 4 for a detailed discussion). There are some good grounds for arguing though that citation rates should be assessed leaving out self-citations – although as Chapter 4 shows there are also strong arguments the other way as well. Cumulatively these effects are more than enough for us to emphasise that no worthwhile comparisons of citation rates or scores achieved by different academics can be made across the major discipline groups recorded in Figure 1.1 The nature of an academic subject, the ways in which it is set up to generate different kinds of publications, and how practices relating to citation and literature reviews have developed over time, are all far too distinctive across major subject groups to make inter-group comparisons legitimate or useful.

Figure 1.2: Differences in the average aggregate citation rates between major groups of disciplines, (that is, total citations divided by number of publications)


Source: Centre for Science and Technology Studies (2007)

Looking in more detail at the detailed variations across individual social science disciplines in citation rates, Figure 1.2 shows that the rates vary from just under a third in psychology (half of which we consider as belonging to the STEM disciplines), down to just over a fifth in economics and political science. In terms of not blowing your own trumpet, these seem to be the most austere disciplines outside the humanities. They also appear to be disciplines where cumulative work by a single team or research laboratory on developing ideas, methods and approaches distinctive to their lab or university plays least role in developing knowledge. Whatever the reason, in all the fields where self-citation is below a quarter in Figure 1.2, it seems there is scope for academics and researchers to be more generous with self-citations. There is also some preliminary research work suggesting that perhaps authors who self-cite, also get cited more by other people than those who are too puritanical in approach.

Lastly by way of introduction, it is important to notice that key bibliometrics and citation tools initially developed in America and some of the, notably the ISI, continue to have a strong built-in orientation (or bias, depending on your viewpoint), towards English language publications. All of the citations tracking systems have begun to diversify in the last decade, but progress has been fairly slow, especially in ISI. Authors who publish exclusively in English will have the most comprehensive citations information. Citations for authors who publish both in English and in other languages are likely to be seriously under-counted on the non-English side. And authors publishing exclusively in a non-English language will be the most under-represented of all.

 

1.2 Academic careers and the accumulation of citations

 

Citation patterns are strongly linked to academic career development, that is, how far along they are in their career trajectory and which route they are following.  process which takes a long time to get started and to develop. Citations counts for academics are therefore highly attuned to age, gender, size of country, and other demographic variables.

Modern researchers and academics often feel under impossible pressures to perform brilliantly in many different spheres of activity, such as forefront ‘discovery’ research, academic integration of knowledge, teaching, academic citizenship and management roles, and achieving external impacts (see Chapter 5).  These combined expectations cannot all be met by one person in a single time or even a single whole career – and yet they are often melded into a single composite image of what the ‘ideal type’ academic should be able to do. This image is unrealistic and disabling because it takes insufficient account of the contemporary specialization of different and equally important academic career trajectories. Modern academia is decreasingly a lone-scholar occupation.  Working in research and academic teams is increasingly important and enhances the need for role-specialization. Finally, different disciplines vary a great deal across the social sciences in how academic roles are configured and in the mix of roles needed.

Obtaining a doctorate and beginning to generate reputable publications both entail overcoming high peer review barriers. Not everyone with a PhD who wants to stay on in academia can do so, which makes the transition to a first post-doctoral appointment (either in a research role or as temporary lecturer) highly competitive. Later on, transitioning from a researcher funded on ‘short’ project budgets, or from teaching fellowships or temporary/junior appointments, to becoming a tenured member of an academic department is also difficult. Being able to generate publications at an early stage- despite the many other demands on young academics- is often crucial for successfully transitioning to a long-run academic career. When academic researchers are in their late 20s and early 30s, and still building up their research skills and competencies, it often takes time for them to produce their first publications.

Once the turmoil of getting onto a tenure track is passed many academics enter into their most innovative and productive stage of new research work in their 30s and 40s, especially in technical or mathematically-based subjects. In this period publications become more frequent as academics become more experienced and formulate better ‘standard operating procedures’ for completing research and publishing outputs. As authors become better known their citations accumulate and their annual rate of citation normally tends to increase. These citations may either tend to reach a ‘steady state’ or plateau, or they may continue to grow rapidly or incrementally, often responding to how far the research community sees their work as successful, reputable and innovative.

As in many other professions, in their late 40s, 50s and early 60s many academics move into more integrative or managerial roles. The most administratively competent or interested senior staff may end up running laboratories, departments or serving in university roles. The more academically orientated senior staff in many disciplines also tend to succeed better in securing funding, perhaps becoming a ‘grants entrepreneur’ and running large-scale research projects. Authors less involved in research teams also often edit journals and co-ordinate academic networks. Senior staff tend to undertake more applied work which usually earns them greater recognition and hence more impacts outside academia itself. The cumulative effect of these changing roles is that senior academics’ research-frontier journal outputs may decrease. At this time academics in most social science disciplines tend to write more books or book chapters, and in many disciplines they continue to play more of a research-leader role in joint articles. Senior academics are also generally better networked and able to draw on even more experience to formulate problems. They have generally established channels of influence and their outputs tends to get more attention in their disciplines. For these reasons their annual rates of citations tend to stay high or keep growing at this career stage. Beyond retirement annual citation counts tend to reduce, as academics are not as active in professional networks as before.

For a minority of academic ‘stars’, however, citations per year may still increase rapidly in their late career for several reasons. Their mature works may achieve wide recognition, often because they have strong integrative effects within a discipline; or their earlier work may acquire ‘timeless’ or ‘standard reference’ status and thus continue to be cited despite being long-published (also guaranteeing close attention to their later work). They may  undertake more applied work that acquires wider influence beyond the academy.

Figure 1.3: Hypothetical citations profiles over time for three main types of publication

These key demographic factors interact with the characteristic pattern of ‘normal’ citations, shown in Figure 1.3. Usually there is an initial lag in the recognition or take-up of published articles of around a year or so and longer for a book. This is followed by a higher-intensity citing period, generally from one to four years after publication in the physical and social science disciplines, perhaps longer in the humanities. This occurs when the work is first widely communicated to a research community and in some way shapes the research forefront – which are the optimal conditions for being cited. Being cited in one place will also create a smaller ‘multiplier’ effect for other current authors to cite the piece. After this peak period passes, however, journal articles will generally drop out of regular sight. They will subsequently be found usually by authors conducting literature reviews and searching with appropriate keywords. Similarly, new books will feature prominently in publishers’ catalogues in their first year, less prominently for between one to three years after that (with research monographs getting least coverage in later years), and then cease to be mentioned. After initial worldwide sales to main university libraries have been exhausted, monographs may only be findable by people searching library catalogues, Google Books, Amazon or the internet. But books that achieve sales to students and professional audiences may be publicised for somewhat longer.

Figure 1.4: Normal and extended citations profiles for individual piece of research

For individual publications we see a three-part pattern of influence, shown in Figure 1.4 – with an initial lag period for recognition, a core ‘pulse’ of citations in the optimal years (usually 2 to 5), and then a ‘tail’ of citations. For regular journal articles this will tend to decline very steeply. The tail may be rather longer for books, especially in the ‘soft’ social sciences, for example, communication and media studies.

 

Different disciplines now vary a good deal in citation patterns across working papers and published journal articles. In political science, for instance, working papers have little currency and journal publication is the key stimulus. Most researchers here also seem to still use ISI and older databases for searches. However, in economics working papers are more important, partly because it may take 3 to 3.5 years to get papers published in key journals. Hence there is a ‘two pulse’ model as in Figure 1.5 with working papers achieving impacts quickly, but subsequently ceasing to be cited when the fully revised version is published as journal article after two or three years. Some prestigious working papers series (such as those of the National Bureau for Economic Research in the US) achieve wide currency as soon as they are issued, along with papers from some major economics profession conferences. Researchers will normally cite the paper in one but not both of the two core versions.

A few of an author’s publications may break out of the ‘normal’ pattern for journal articles and research monographs of peak and decline, and instead will achieve relatively higher levels of continuing references.

  • an ‘enduring’ piece of research still has a falling citations profile over time, but one falling more gently and stretching beyond 5 years;
  • a ‘standard reference’ in a discipline or sub-discipline will be distinguished by having a stable tail of continuing citations below its initial peak, but which does not thereafter decline for an extended period, perhaps as long as 10 to 12 years. Standard references may benefit from being a ‘first-in-field’ treatment; or they may have strong multiplier effects; or they may just be located in slower-moving or less popular parts of a discipline. Finally,
  • ‘classic’ pieces of research can be distinguished because their over-time annual citations volume tends to expand for the same extended period (say 10 to 12 years, perhaps even beyond that in some cases).

How do these citation patterns affect the over-time profile of individual researchers and academics? Figure 1.4 shows three fairly widespread patterns. Researchers whose output are episodic and separated by longer periods of time may have an over-time profile of small numbers of citations pulsed around the episodes of their work coming out. Academics that become better established, and can crank up a reasonable rate of publications and maintain that regularly, thereby benefiting from the accumulation of citations for different pieces of work. Their annual citation rate will hence grow steadily in their early career years, reach a ‘plateau’ level fairly soon (perhaps most usually in their mid to late 30s) and broadly maintain that level (perhaps with a few ups and downs) until retirement. Finally, the most successful academics will not only benefit from the short-term blips of citations for their regular work, but will add layers of continuing citations from items that become enduring, standard references or classic references, perhaps especially books in the ‘softer’ social sciences. Pieces of work that achieve these longer tails, plus a more intense pace of research outputs in mid-career years, both help successful academics to achieve annual citation rates that usually grow over long time periods, along with their seniority. Here retirement may not have immediate effects on reducing cumulative citations.

 

1.3 Different career trajectories and publications profiles

There are many different routes for academic career trajectories, which could be characterised in a large variety of ways. For our purposes here, Figure 1.5 shows that a key branching point occurs between two paths, one that is research-predominant (conceivably research-only in some subjects), and the other which is a teaching plus research track.

Figure 1.5: The research-intensive and teaching-based pathways in academia

This divergence tends to occur early on during someone’s doctoral work. The factors that incline people one way or the other at this and later stages are always complex, and so to summarise may always be to over-simplify. But an early factor that often seems to set people onto one or the other of these tracks concerns the extent to which their doctorate is undertaken as part of a large research team and in a university context that plugs them into strong networks in other universities, perhaps internationally. PhD students who are well plugged-in seem to be also more likely to adopt topics and approaches that lead more to research-track progression. They may also commit more strongly to attending professional conferences and do more ‘fashionable’ or forefront work.

By contrast, students who work on their doctorate in smaller departments or more on their own, with relatively lesser supervisory or peer-group support, tend to focus more on topics that may not lead easily to winning posts in research-intensive institutions at a later state. They often also invest more in developing their teaching capacities early on, and locating their futures within more teaching-orientated universities.

Research-dominated and research-only careers are far more feasible in the physical science subjects (including medicine), engineering, technology and mathematics (hereafter termed the STEM subjects) than in social science. One key factor behind this already operates strongly at the next key stage where people make the key transition from working on a PhD to getting a post-doc position. The latter are usually concentrated in research-intensive universities by patterns of government funding for the STEM subjects, and hence the availability of such posts varies sharply across disciplines.

Figure 1.6: The growth in the number of US PhD holders who have ever held post-doc positions, by discipline groups from 1972 to 2006

Figure 1.6 shows that in American universities three in every five PhD holders in the life sciences have held a post-doc position. By contrast, in the social sciences the proportion is half this level, at three in every ten. However, in the social sciences this proportion has grown slowly over three decades, from less than two in ten in the 1970s – more or less keeping pace with other disciplines where post-docs have been increasingly common, especially engineering and computer science and mathematics. In the physical sciences the proportion of social science PhDs with post-doc experience has oscillated quite sharply with changes in the economy or the availability of funding. By contrast, at least the experience of post-docs in the social sciences has been very steady over time.

At the next stage in the STEM subjects also there is additionally a regular supply of contract research positions that can be alternatives to post-docs, and involve working in a junior role on larger scientific projects, funded from ‘short money’ budgets in university departments. There are also often extensive research positions in business labs and technical institutes, from which it is still feasible to return to an academic pathway. not all academics do a lot of teaching. As a result many researchers are able to develop viable, research-only career routes, where their teaching outputs are minimal or zero.

Figure 1.7: Numbers of UK academic teaching and research staff, and sources of funding, by discipline group in 2005-06

All disciplinesSocial sciences and humanitiesScience, technology, engineering and mathsCreative, arts and design
All academic teaching and research staff160,00059,80088,00012,400
Staff who only do research36,8005,50030,800600
Percentage of all staff who only do research239355

Source: HESA statistics 2005-06; LSE Public Policy Group (2008, Figure 1.2).

In the UK Figure 1.7 shows that more than a third of UK university academics hold research-only posts in the STEM disciplines – reflecting the concentration of five sixths of dedicated government research funding on STEM subjects, plus the additional commitment of corporate research and development monies. By contrast, only one in ten professional social scientists in higher education has a research-only job. In particular, the vast bulk of social science academics undertake both research and teaching throughout their careers.

Alternatively in Figure 1.5 young academics may get appointed to a time-limited teaching contract, as a junior teaching fellow or on a short-term appointment as a lecturer or assistant professor, often in universities most orientated to undergraduate teaching. The scale of temporary appointments involved here has mushroomed in recent decades, as universities have run down the proportion of their staff who are full-time and tenured faculty and increased their use of part-time and non-tenured teachers. These developments match similar changes in a wide range of business and government organizations towards more ‘flexibilization’ of staff by organizations, with individuals more commonly having a ‘portfolio’ career path with multiple components, rather than lifetime careers with a single employer.

On an individual level, getting into one track or another often makes a large difference to the probabilities of subsequently publishing, but of course it is never decisive or fully determinant. At later stages people can shift between tracks, with initially teaching-track academics who undertake excellent research tending to move into more research-intensive universities over time. The bifurcation between ‘research-intensive’ universities, laboratories and departments and departments in predominantly teaching-orientated departments is important, but it is also not complete. Even in the most research-intensive institutions some members of departments will be more ‘research active’ than others, and some will be more teaching-orientated, especially taken across long careers. And even in mainly teaching-based departments, a lot of good research gets undertaken – as the 2008 research assessment exercise in the UK demonstrated. Even in a bureaucratic exercise weighted to legitimizing existing funding distributions none the less showed multiple ‘islands of excellence’ in smaller universities and departments. In some university systems, however, the separation of research-lead and teaching-orientated universities has stronger consequences. In the US some evidence suggests that staff from the many non-PhD departments in the social sciences generate only around a sixth of journal articles in their discipline. The situation is less stark in the UK and Europe, where almost all universities will claim to run PhD programmes in most subjects, but there is still a gap.

So although we acknowledge that the pathways in Figure 1.5 are approximations only, it is still useful exercise to use them as a framework to discuss how individuals characteristically develop some of the main dimensions of their academic activity.

 

Figure 1.8: A ‘balanced scorecard’ for assessing academic achievement

To think about these dimensions we use a fairly simple conceptual schema known as a ‘balanced score card’. This is an approach that developed in business and government as a way of coping with the complexity of assessing an organization’s overall performance. Our schema, shown in Figure 1.8, charts an academic’s profile as low, medium or high when moving out from the centre along each of the six dimensions shown. We begin by exploring the earliest-developed capabilities, coloured blue in the Figure, then move to those shown in green and finally cover the red dimensions.

Research skills and competence are in many ways the first developed aspect of any academic’s profile, since everyone entering the profession now must complete a doctorate. In the physical and social sciences this means that they master an increasing range of methods and skills in an increasingly systematised and professionalised way. In the humanities research capabilities are more varied, typically involving more stress on theoretical and thematic ideas development, and on archival or literary-based methods fro analysing texts. Research competences also typically continue to grow strongly in post-doctoral and early teaching posts. But in principle this is an area where researchers can keep pushing their competences outwards throughout their careers, especially at points where they change topics, or sub-fields, or the direction of their work.

Authoring capabilities are normally the second dimension that academics develop early on, usually somewhat lagging behind research skills. In the physical sciences especially, the tradition has been to undertake series of experiments first (for say two or three years) and then to ‘write up’ extensively only at the end of the doctoral period. In many (but not all) STEM subjects, composing text this final text for submission is also often still done in a restrictive technical structure and format. In the social sciences and especially in the humanities, however, it is more normal for people to treat writing as ‘constitutive’ of their thinking, and hence to write chapters as they go along (Dunleavy, 2003, 2009). In ‘soft’ subjects students write a ‘big book’ thesis where how a researcher’s authoring skills shape up early on often determines to a large extent (say 30 to 50 per cent) how successful their PhD is and whether they are able to generate early journal articles (Dunleavy, 2003). In other more technical social sciences (like economics) most students now complete a different ‘papers model’ PhD, which is a shorter text, but where the authoring and presentation standards are higher and where the three or four component chapters must attain a higher, ‘publishable’ quality (Dunleavy, 2011). Whatever pattern is followed, during the later years of their doctoral work anyone entering modern academia must begin to strongly develop the (admittedly often strange or off-putting) forms of professional writing used in each discipline. In all subjects following the ‘big book’ model, a PhD thesis can often be (one of) the longest piece of sustained writing that a researcher completes across their academic career.

Teaching capabilities are the third dimension that would-be academics start developing in the middle to later years of their doctorate, when they begin teaching classes and seminars, and perhaps giving a few lectures. In some subjects PhD students often take on course administration tasks, and even examining responsibilities also, especially in the US and Europe. Nowadays most PhD students in the UK complete a more structured programme for developing their teaching capacities and skills, with certification linked to the Higher Education Academy, a body that inter alia provides assurance to future university employers of their basic competence.

However, the most critical stage in the expansion of teaching skills occurs when new lecturers or assistant professors start work full time and begin to cope with a full teaching load, often initially in temporary or time-limited posts. Their employing university will normally provide formal induction processes designed to enhance their capabilities, and their department may require completion of a formal certification as a competent teacher (especially for tenure track posts). Beyond this beginning stage, teaching capabilities generally take many years to develop as academics’ experience of different types of courses and student groups grows, from undergraduates through masters courses and extends to PhD teaching and supervision. So far, academics have generally been exempt from the requirements to periodically re-certify their professional competence that are common in other professions, such as medicine and law. However, in modern universities student feedback scores provide a ceaseless commentary on teachers’ success and some spur to continuing improvement.

Management capabilities for academics and researchers generally are acquired informally as their career develops and they assume senior positions. Academic management roles relate chiefly either to running combined teaching and research departments or in the STEM subjects to full-time posts organizing research units and labs. The implication of Figure 1.7 is that in the social sciences perhaps 90 per cent of management tasks relate to conventional academic departments, while running research units and labs is either much less common or more of a part-time commitments. Indeed, in many universities and in ‘softer’ social sciences with less team-effort in the research process, academic management is almost synonymous with departmental staffing and administration issues.

As in most other serious professions, management capabilities tend to be developed as people become older and more experienced, normally in their late 30s through to their 50s (for someone entering academia in their late 20s or early 30s).  Universities have some rudimentary training for heads of departments, but these capabilities are primarily inculcated across the sector in a rather amateurish way – by ‘socializing’ academics into administration issues piecemeal. The core process here involves the parcelling out of numerous administrative chores, along with broader ‘departmental citizenship’ tasks and the job of representing the department on numerous faculty or university committees. Younger academics get to do the more boring or tedious chores here, and with age and experience gravitate to more consequential or outwards-looking roles. The process may seem rather random and disorganised, and academics often spent inordinate amounts of time bewailing having to handle a quota of administrative and bureaucratic tasks (‘whinging’). However, universities are very unlike what Henry Mintzberg calls ‘machine bureaucracies’, by which he means the classic forms of administration of firms and government bureaucracies analysed by Max Weber.  Instead universities are classic ‘professional bureaucracies’ with a quite different internal structure as described in Figure 1.9.

Figure 1.9: The key differences between universities as ‘professional bureaucracies’ and Weberian or ‘machine bureaucracies’ (such as government agencies or some large private corporations)

Notes: Each form of bureaucracy includes five elements, but their relative sizes, roles and powers vary a good deal across the two types. The ‘strategic apex’ covers the controlling decision-makers and their immediate support staffs. The ‘middle line’ covers the routing of resources to production and the supervision of what gets done. The ‘operating core’ is the part of the organization that implements production or carries out the core ‘mission’ of the organization. ‘Support services’ are things that support the organization’s main mission but are not part of it directly (and so could be outsourced in the modern era). The ‘technostructure’ is the part of the organization that innovates, designs new products and pushes forward organizational efficiency.

The rationale for universities’ apparently unusual approach to organizational management has always been to maintain a close control of all university politics, decision-making and management by their academic departments – what Mintzberg calls their ‘operating core’, the part of any organization at the heart of its mission. Like other professional bureaucracies, universities are politically dominated by their ‘operating core’ – so that their professional academic staffs decide their policies in a collegial manner. Compared to machine bureaucracies, universities have minimal ‘middle managements’ and a curiously undeveloped ‘strategic apex’, because the academics insist on retaining so much control in their hands. They also have big support services (covering functions such as libraries, IT services, collecting student fees and research grants, and running catering facilities and halls of residence). But however large-scale they become, these operations are kept in a very subordinate role to the dominant professional group, namely the academics.

Professional bureaucracies also have very slender innovation, improvement or product-development specialist units (called the ‘technostructure’ in Mintzberg’s terms). The vast bulk of this work is instead done by the academic departments themselves. In research-intensive areas, the effective organizational management of labs and specialist units requires very high (post-PhD) levels of context-specific information, expertise and understanding, as well as more generic leadership and management skills and capabilities. Thus universities depart in many key respects from modern machine bureaucracy paradigms in business and the private sector (Roberts, 2004) or in the government and public sectors. The apparently haphazard socialization of academics into management roles plays a key part in maintaining all these features. But, just as in other large organizations, managerial capacities still form a key part of the burdens of seniority.

Networking is an academic skill that develops over time and is clearly linked to research in several dimensions. At its most basic, the ability to work in teams of two, three or a few co-researchers and co-authors is an important influence on the quality and type of research that any academic can undertake. Modern social science is more specialised than in the past, yet co-author teams have remained much smaller here than in the STEM disciplines. Networking and the ability to build teams is also important for winning research grants, itself a key influence upon research productivity – given that most social scientists have continuing teaching obligations, from which grants allow them to be bought out. Academic networking within disciplines but across universities and countries is a key element in broadening academics horizons, keeping researchers in touch with the constantly-moving research frontier, and up-to-date with recent substantive and methodological developments.

Networking within universities across disciplines is often a key influence on inter-disciplinary research, as is academics’ ability to engage in ‘bridging scholarship’ that works across fields and helps develop meta-theories and intellectual waves – both of which influence external impacts (see Chapter 5). Finally, networking to external actors is a key element in fund-raising for research from non-foundation and non-government sources, and it forms a key element in academics achieving external impacts at later stages of their careers.

Celebrity is the final dimension in Figure 1.8, and at first sight this label may seem an odd one to choose. Do not academic capabilities and professional virtues stand in acute contrast to the ungrounded, ‘famous for being famous’ quality of celebrity in contemporary media or popular culture? Of course, as a result of peer review academic reputations are normally grounded in more solid and well-attested achievements. But it is also clear that the distribution of fame and knowledge of their work and arguments across academics is highly uneven. Some excellent academics are little known, and some of those who become well-known are not necessarily strong figures in intellectual terms.

What shapes academic celebrity? In a famous analysis of ‘public intellectuals’, Regis Debray (1981) argued that there have been three phases of development in their characteristic origins and roles since the last quarter of the nineteenth century. The first was the age of universities’ pre-eminence, from the 1860s to the early twentieth century. The second was an era dominated by writers and literary figures, from the 1900s to the 1950s. The third is the age of public intellectuals as media-savvy celebrities, whose reputation depends far most closely on their ability to project and convince via the mass media. This still-current period dates more or less from the advent of pervasive television coverage in the mid 1960s onwards. Arguably Debray’s analysis is overly orientated to a restrictive French concept of public intellectuals, and it neglects the enduring role of science-based intellectuals, who remain resolutely university-grounded. Yet the growth of popular science books and media productions, and of science/technology-watching magazines and newspaper columns, has also contributed to the emergence of ‘celebrity scientists’.

The apparatus of achieving academic ‘celebrity’ has also drastically simplified and been democratised in the digital era, so that internet mechanisms are now reasonably decisive in conditioning someone’s renown. Counting an individual academic’s cites in Google Scholar or Google Books, is a once-specialist activity that can now be easily (almost instantly) undertaken by anyone. Their prominence in ISI ratings or Scopus is a bit more tricky because of the access costs involved, but even these older, paid-for databases should be equally instantly available to staff and students across the university network.

So what was once vague or requiring expert judgement has now become simpler. We can index and measure someone’s prominence on the ‘celebrity’ dimension perhaps more easily than almost any other. In an increasingly globalised academic community, the importance of academics’ and researchers’ wider reputation in attracting attention to their work has never been greater. Celebrity has hugely increased in importance relative to networking interactions. Whereas once academics relied on people knowing them and their work personally in order to gain citations from other academics, now what matters is how easy it is to find someone’s work – and how many versions of it there are out there in different channels to be picked up and noticed by other academics and researchers.

Similarly, contra Debray, academics’ dependence upon mass media intermediaries to reach any audience beyond their immediate discipline has arguably reduced in an era where full academic works can be accessed through the internet at the click of a button. A whole series of developments have recently coalesced to begin far-reaching changes in the inter-relationship of academic work and wider societal development in advanced industrial societies including:

  • Google’s push to ‘organise the world’s information’, especially via its Scholar and Books operations;
  • the growth of free research depositories for academic materials, making them much more accessible to non-professionals;
  • improvements in professional communication with the public in the physical sciences and (after a long lag) the social sciences; and
  • the emergence of many think tanks, a burgeoning industrial and professional consultancy sector, and numerous NGOs and specialist media interested in debating and processing much more specialist themes (see Chapters 5 and 6).

These changes have occurred rapidly in the last two decades, but in many ways they have only just begun and they have a long way further to run.

Celebrity has also begun to change the ways in which government, business and to a lesser degree actors in civil society gain access to academic expertise. As late as the 1980s officials in government departments especially could afford to maintain costly, long-run personal networks of contacts that formed their gateways into seeking external expertise when needed. Since the age of ‘new public management’ and the subsequent austerity period following the 2008-10 financial crash in many advanced industrial societies, government’s apparatus has been pared back. Now when they need academic expertise, UK civil servants told us for this research that they go on Google and search digitally like anyone else, and as their American counterparts have been doing for a decade or more. In many STEM disciplines large business corporations close to particular academic discipline areas still operate networks based heavily around personal contacts, as do business schools in most countries and some increasingly specialist public policy schools in the US, Europe and elsewhere. But increasingly academic celebrity rather than personal contacts has become the currency by which the media initially and other sections of society form a view of the debates and knowledge-terrain inside disciplines.

How should we weight or compare individuals’ achievements on the six dimensions in Figure 1.8. The whole rationale of such scorecards is that organizations (and here individuals) need to do many different things at once, all of which need to be kept in view for an accurate assessment of their progress to be made. Thus, a firm that makes short-run profits by taking big risks or neglecting to invest in its talented-development or business infrastructures is not a good investment. And nor is a government bureaucracy doing well if it saves money by worsening the standards of services it delivers to citizens or cuts corners on consultations or rule of law principles. The rationale for using a balanced scorecard approach to assess academics is very similar. In the same way, the six dimensions in Figure 1.8 are all important in some combination for all kinds of university professionals.

But this is not to say that any given person can or should be expected to perform excellently on all these dimensions. A disabling paradigm of the ‘perfect academic’, who is good at all these things simultaneously – a great researcher, author, teacher, manager, networker and celebrated disseminator of knowledge – often lurks pervasively in the culture of higher education and academic disciplines. It shows up strongly in appointment, promotion and appraisal discussions and it is pervasive in the pages of most universities’ human resources manuals. This mythical image of an omni-competent academic is also powerfully codified by government bodies conducting research audits (like the UK’s Research Excellence Framework) and by government or foundation grant-giving bodies demanding ‘impact’ and ‘dissemination’ from those to whom they dispense funding. Yet our argument here is that no one can be simultaneously good at all six of the dimensions we have reviewed. Instead most academic career tracks involve people in specializing to a considerable degree, and thus ending up with a configuration of capabilities that will differ significantly from those of other academic professionals who choose alternative career routes. To explore what this means in practice, we follow through in more detail how people’s capabilities develop at the key stages in the two trajectories shown in Figure 1.5, beginning with the research track sequence of roles.

Figure 1.10:

Here Figure 1 .10 suggests that PhD students are likely to have their best-developed capabilities on the research dimension, where they should score medium, because they are still learning the craft of research at this stage. At the same time they will have to achieve at least a basic competency in authoring (to communicate their findings), in teaching (which even research-track people must usually do at this stage for pecuniary and career-development reasons), and in networking (essential if they are to have a decent sense of where the research frontier is and of the requirements for career progression). Most PhD students will not have developed even low managerial capacities, nor will they normally rate any level of celebrity.

Moving on to being a young academic in the research-track is a time when people’s capacities improve in many dimensions at once, with changes shown by the arrows. Thus the second chart in Figure 1.10 shows individuals growing their research capabilities from medium to high; improving their teaching, authoring and networking capabilities from low to medium; and establishing low capabilities in managing and in terms of celebrity and citation scores. Achieving such a multi-dimensional improvement is an extraordinarily demanding thing to do, and younger staff can expect to work many hours a week to get it done, perhaps in a way that is not sustainable over the long term.

At a senior academic stage in the research track, Figure 1.10 suggests three possible patterns of development, each much more sustainable over the long term once the first burst of career-establishing effort has occurred:

1) The senior all-round scholar profile involves maintaining a high level of research capabilities, while expanding teaching and authoring performance to a high level, and growing managerial capacities to a medium level. The costs of achieving this transition is often that the researcher in question does not become any better connected in professional networks and that their ‘celebrity’ level remains low, with publications staying resolutely academic and discipline-bound.

2) The ‘research obsessive’ profile here captures senior academics whose central focus demonstrates a continuing commitment to research allied with passion for their discipline or sub-discipline. Scholars here become more specialist and focused in their interests and so invest heavily in expanding their networks, especially internationally as their seniority rises. This emphasis fits well with the ‘lone scholar’ mode of research in the social sciences and humanities, and is perhaps less common in the physical sciences where teamwork is fundamental and the financial and time costs of research are high. Research obsessives may achieve a continuing research profile only at the expense of not much expanding their management capabilities (they shun all administration) and remaining little known outside their sub-field (so their celebrity score remains low). They are also not known for being outstanding teachers.

3) By contrast the ‘research grants entrepreneur’ denotes a crucial role in areas of research like the STEM disciplines where assembling and funding teams of researchers is vital for achieving advances. Here academics tend to withdraw from teaching to focus on leading a research unit, and they may also do markedly less authoring for lack of time and because of the specialization of roles within the research team. Instead grants-entrepreneurs maintain their research capacities at high, but also expand their management and networking capacities to high. To help win grants and to tap wider resources beyond grant-funding foundations or government bodies, they must also become at least moderately well-known, expanding their celebrity capacity to medium. Grants entrepreneurs, of course, rely on junior researchers to undertake virtually time-consuming primary research, and often to write up the first drafts of papers, with their contribution being in intellectual leadership, managing team members, providing a fount of accumulated experiences for the team to draw on, and securing a continuing funding stream.

Figure 1.11:

Turning to teaching track academics, Figure 1.11 shows that the only salient difference at PhD level is that the student here is not networked. Indeed the context of many students’ doctoral work remains very closely bound by what is going on in their home university department, and perhaps a little beyond. PhD students here have a medium research capacity (because they are still learning), and a low capacity in authoring and in teaching, where the roles for PhD students are inherently rather limiting.

Once people on more of a teaching-track make the transition to being full-time young academics with tenure, Figure 1.11 suggests that they invest heavily in boosting their teaching capabilities from low to high, in expanding their management and administrative capabilities from zero to medium, and in improving their authoring to a medium rating. Not surprisingly, achieving this degree of change leaves little time for expanding research capabilities and methodological skills, which may stay at medium as a result. In terms of celebrity younger teaching track academics may remain very low profile, since their publications are new. However, most people at this stage will expand their networking capabilities to at least a low level.

For senior people in this track there are three possibilities for further development:

1) The senior all-round academic profile here is exactly the same as that already discussed in Figure 1.10. What is different in Figure 1.11, however, are the arrows showing the degree of change from the young teaching track profile. Teaching capabilities stays at high and management at medium, but as they become more experienced with successive projects and writing articles and books senior teaching-track academics invest heavily in expanding both their research and authoring capabilities to high. The accumulation of publications also expands their celebrity from zero to low, and their networking capabilities from low to medium.

2) Some senior teaching-track staff specialise instead in academic management roles, running departments, and often moving on to undertake university roles as well. While keeping their teaching capabilities at high, and their research and authoring at medium, they invest in moving their management capacities from medium to high, which absorbs a lot of time. Broadening their management roles also tend to expand their networking capacity, while their accumulation of publications and citations expands their celebrity from zero to low.

3) Finally some senior academics in fields where lone scholar research prevails (as in many humanities and ‘soft’ social sciences) may transition to a ‘pop academic’ profile, as may some individual expositors in areas more dominated by research-team work (such as ‘popular science’ expositors). Here the academic tends to withdraw from teaching and strongly avoids all administration (so that their capacities on both dimensions may decrease). Their research capacity stays stable (at medium) but they specialise strongly in achieving excellent authoring skills, which move to high. Other expository skills, such as lecturing, designing media programmes and expounding in person on TV also move to high. Well-known academics will invest time and effort in becoming strongly networked (where their score improves from low to high) and household names in the media and externally (where their score improves from zero to high). Authors here may achieve high level of citations for books that expand public understanding of their discipline, but many also undertake important scholarship in a more ‘integrative’ vein focusing less on discovery research and more on thematic or theoretical understandings. And academics in this stream may often have high overall external impacts also, parlaying their celebrity into influence also with businesses or governments. But in other respects this still remains a somewhat risky choice of career-turn, as James Boyle noted:

For those in my profession, being readable is a dangerous goal. You have never heard true condescension until you have heard academics pronounce the word ‘populariser’.

Trying to categorise diverse academic career pathways into just a few types (as we have here), risks over-simplifying a complex picture. Yet we believe that it is worthwhile to do so in order to stress that people at different stages on different career paths are likely to have quite distinct profiles of citations within academia, and quite different impacts outside the higher education system itself. Research track academics, as we have described them here, are likely to fare well in the most conventional, journal-orientated bibliometric systems, such as the ISI Web of Science discussed in the next chapter, whereas teaching track staff are likely to fare better in broader bibliometric systems, such as Google Scholar and Google Books. Younger staff are likely to have slender citations profiles, and senior staff will generally fare better in cumulative citations terms, although their annual rates of citation may not be so different.

 

 

Summary

  1. Citation rates are used as a basis for tracking academic impacts. The shape of citation rates vary widely across academic disciplines.
  2. There are substantial differences in the general rate of citing across disciplines with more cites (including self-cites) being found in the sciences than the social sciences.
  3. The type of output chosen affects citation rates e.g. on average a book will take longer to be referred to but will be cited for longer.
  4. How academics balance their time across the six areas of responsibility will be another important factor in citation rates.

Back to top

Print Friendly