to handle a substantial bodv of evidence in a way that is fair, persuasive and
uncluttered.Because the research report will often be longer than a term paper or
essay,it can be daunting,which can lead somepeople to make the seriousmistake
of puttingoff the writingfor as longaspossible. In addition,the lengthitself poses
problems of organization, structure and coherence that are more acute than in
essays.
Although a conventional academic report is the most common product of
research, there is growing interest in creative expressions of research findings in
the formof broadcasts,narrative,story and film.It is worth recallingthat Balzac's
Comtdie Humaine, Dickens' Hard Times and Zola's the Rougon-Macquart are
works of social analysis that could, especially in Zola's case, be located within
critical theory. In Britain today, writers such as Peter Ackroyd and Pat Barker
show the power that fictionhas authenticallyto represent a past, and to do more
besides. These are interesting, exciting developments, but we will not explore
them here, concentratinginstead on mainstream report writing.
Starting Writing
This sectionis influencedby two books. The first,Boice's report (1992)of a well-
thought-out programmeto support new members of academic staff, may seemto
be a&odd inspiration. Undergraduate and postgraduate students writing reports
are not in the samebusiness as full-timeacademicstaff: studentsare novices, staff
are experts; students strain to write, but lecturers' fingersfly acrossthe keyboard,
with neat phrases smoothly appearing and complex concevts rolling across the
pages. Or so those who are not members of academic staffmight think.
One of the striking things about Boice's research is that he showed how
difficult people find writing, even when they have already produced their PhD- A
theses. Investigating firther, he found that the academics who were the most
productivewriters were not thosewho waiteduntil they had cleardays when they
could 'binge' on writing. Those days disappear under tides of other tasks. The
productive writers were those who tried to write a little regularly, or even daily.
That is our firstmessage: write regularlyand throughoutthe project. Do not try
to write the report in one go. It is better to draft an introduction (which will
certainlybe rewritten later) very early in the research.Therewas advice on doing
the literature review in Chapter 4, which should be referred to as the writing
continues. The suggestion here is that the review is best written up while the
research and thinking are in progress, again in the knowledge that it will be
reshapedwhen the research is done. Likewise,it is a good ideato writethe section
on the research methods when decisions are made about the research design and
whilepiloting takes place. Field notes and memos you wroteto yourselfwill often
provide draft material for incorporationin the report.
That is one of the strongest themes in Becker's book (1986), which is the
second that influences this section. Keep drafting, editing, shaping, changing.
That is the best way of beating writer's block. The trick is to write sometlzing,if
only as a way of helping you to get to something better. Remember that not
everything that is written is 'good' writing (whatever that is). This chapter bears
few traces of draft 1, although it is recognizably close to drafts 2 and 3. Without
the experimentation of draft 1,drafts 2 and 3 would not have been possible. Key
meaningsemerged in the course of doingthat first draft, and if many of its devices
and phrases have gone, the ideas remain.
However, it is important to be able to look later at this quick-writingwith a
fresh, appraising eye so as to be able to edit and revise it. Three ways of bringing
a fresh eye to your draft are:
Re-read what you have written with a style book beside you.
Put the draft away for a couple of weeks or so, and come back to it as an
unfamiliarpiece of writing.
Get hiends to look at it.
The first strategy is good for improving style, but it is not much help with
evaluating the ideas, evidence and structure. The second is a good strategy if the
time is available.For example, the first draft of this sectionwas laid aside for five
weeks, which meant that it could be re-read with more detachment and with a
sharper critical intent. The third strategy is the best, although you need to be
prepared for criticism and suggestionsthat may not be entirelywelcome.
Most ofus have feelingsofunease, unworthinessand riskwhenwe showothers
our drafts. These feelings are well described by Pamela Richards in Becker's

Writing the report 177
more than the summary has a full understanding of what the report is
The summary is often put in bullet point form with cross-references
n each point and the body of the report.
difference between writing for academic and other audiences can be
htedby looking at the way referencesare given.Academic writing is full-
ial that no one is going to be able to locate, let alone wish to read. We all
fall short of the ideal of givingreferences only when:
are summarizingdistinctiveviews or findings.
direct readers to useful sources of further information.
writing a report for a non-academic audience, it is advisable to keep
it takes little space to write a short (up to 50 words) commentary on each one,
summarizing what it says and indicating ~ t ssignificance in the context of the
Criteria for judging research reports
Near the beginning of good programmes of study academic staff will set out
the criteria by which research reports are to be judged. Better still, they may
spend time working through the criteria, perhaps by having students grade one
or two anonymous reports, so that the criteria are not just availablebut are also
understood. While criteria will vary by institution, subject and level of study,
there are some things that are generallyvalued and others that signify an inferior
beliefs that describe bad writing. It sometimes makes no difference when a tutor report. Box 12.1 offers a view of these characteristics.
makes it clear that long words are not better than short ones and that making the
meaning clear is the first task of the writer. Other advice that frequently gets
ignoredls that writingneeds to be 'reader-friendly', so that the readerknows why Organizing the report
something is being covered in a certainway and where the argument is going.
If the report is fo? practitioners and other interested people, bear in mind that It is easiest to organize the report around one of the conventions. Scientific
they will not usually like a dull academicstyle and will complainthat academics' reports, such as those found in most psychology journals, tend to have a very
technical terms baffle them: that the report is littered with jargon. One way of formal structure, which may be quite suitable for the presentation of survey
checking ki;h<tf& your writing is suitable for a lay audienceis to give it to friends findingsto aud~enceswho expectreportsto follow that form. Conventionsare less
who do not know your subject, let alone your research. Listen very carefully to binding when writing ~nother disciplines and for other audiences.
their comments. If they have difficulty with any part, the best advice is to assume The difference between a good and an acceptable report does not usually
that you have not written clearly enoughfor that audience. Change it. depend on how good the research was. The coherence of the report and the

Unresolved issues are explored and directions for further research are
indicated.
Realistic suggestionsfor action are made where this is appropriate.
There are signs of originalityor of intellectual excitement.
The report makes a useful and secure contribution to understanding of
the field.
Acceptable report
An agceptable report will have many of the characteristics of an
outstandingreport but not all of them. The last criterion is unlikely to have
been met. Retvriting each criterion at the acceptable level produces the
followingprofile.
xbq,pqblem or issue is stated.
The findingsor argumentmay or may not be stated early in the report.
Where they are, it is not easy to see in the body of the report that they are
the best conclusionsfrom the investigation.
Links are made to other work in the field but they are not sufficiently
full.
The issue of significance('who cares?') gets passing attention.
The research methods are described.
It is often not shown that they are the best ways of investigating the
issues raised in the literaturereview.
Within the limits of time and resources, the research designpromises to
shed some light on the questionsraised in the literaturereview.
continued
Writing the report I 79
The findings are comprehensiblewith some effort by the reader.
The conclusions are not strongly related to the findings and to the
literaturereview.
I Unresolved issues may not be explored and directions for fvnher Iresearch may not be indicated.
There may not be any realistic suggestions for action, even where that
would have been appropriate.
Unacceptable report
The acceptable/unacceptableborderline is very much one of local custom
and practice. In some cases, falling short on one of the criteria of
acceptability is enough to ensure failure. In others, there may need to be
weakness on several counts before a report is regarded as unacceptable.
There is little point, then, in repeating the criteria of acceptability, but
casting them in the language of failure. What follows is a list of some
distinctive features of poor reports. However, some of these featuresare to
be found in work that is otherwise acceptable:
The prose hurts. This may involve poor spelling; the grocer's
apostrophe('the informant's saidthat ...');highfalutinwritingthat fails
('the problematic in this issue is comprised of. ..'); long and pompous
sentences; difficultyin understandingwhat the writer is gettingat.
The literature review is too short. That does not mean that there are
insufficient references; it means that the writer has not really got the
essence of the sources that have been used. In these cases, the review is
often organized as first a summary of this book and then of that article.
It is like reading a shopping list. Good reviews are organized around
issues and themes.
It is hard to understand what the findings are. This section is often
confusedand frequentlyis organizedas a question-by-questionreport of
the findings: far better to present the findings in terms of the questions
posed at the beginning of the report and developed in the literature
review.
Poor reports often seem to produce conclusions like a conjurerpulling
rabbits from a hat, avowing that what has gone before self-evidently
leads to those conclusions. The conclusions are disconnected from the
rest of the report and no amount of insistence that 'it has been shown
that. ..' can hide it. In short, there may be some information in the
report but it is not organized so as to help the reader know what it is
supposedto signify.
References are a mess. Few people are failed on the basis of bogus,
incomplete, or poorly presented references. However, these faults can
make readers suspicious~f the rest of the report.

180 Interviewing for social scientists
Introdu<tion This should say what the problem is, why it is of any interest or
importancewandset out the main conclusions of the research. It may include
or take the form of;an executive summary. Avoid an introduction that says that
certain methods were used to investigate the problem, which led to unspecified
conclusions. That is weak and makes it seem as if you do not know what you are
trying to&ay.J&,eepthis section short, allowing perhaps 5 per cent of the overall
word count for it.
Background This section may not be needed, since the literature review may
effectivelyprovide the background. Where it is necessary to supply some back-
ground that cannot be presented elsewhere, keep it brief and to the point.
Historical backgrounds often contain a lot of information that is not needed in
order to understand what is being investigated or claimed. Remove it.
Literature review Discussed in Chapter 4, this is one of the most importantparts
of the report. It should set out the state of thinking about the problem or issue
that you have investigated; identify gaps in existing knowledge, or flaws and
problems with it; note methods of enquiry that have been used in this area; and
conclude by identifying the key questions that will be studied and the methods
of enquiry that are likely to be productive. This can take 20-25 per cent of the
report,where it is a report for an academicreadership. It can also be a lot shorter.
Research design Show that the methods used were the best of all those that might
have been used (see Chapter 6 for an example). Details of the development
and piloting of interview schedules should be included, ideally showing how
each question or planned prompt was designed to explore an aspect of a research
Writing the report 181
question listed in the previous chapter. In the best reports, there is a clear and
strong link between the design and the literature review. Include details of Sam-
pling and discussions of credibility and generalizability.This is also the place to
give serious attention to something that is frequently side-lined, namely the
analysis of the data. Methods of data analysis should be discussed and there
should be some evaluation of the confidencethat can be placed in the outcomes
of this analysis. Finally, consider the strength of the design overall, if need be,
indicating how more time or resources would have allowed for improvements.
Allow up to 20 per cent of the report for this important section.
Results Saywhat you found. Somewriters liketo interweavereporting the results
and commenting upon them, in which case this section will be combined with
the next, and probably be organized into several shorter chapters. Each one will
addressa key theme in the data,describingwhat was foundand commentingupon
it. Other writers like simply to present the data, perhaps using the end-of-chapter
summaryto indicate the themes that will be pursued in the discussion.Where the
results are presented without commentary, give no more than 30 per cent of the
report over to them.
Discussion This section is one of the most important in a report and the one that
often does more than any other to determine how well it is received. It needs to
appraise the results and their meaning mainly in terms of the themes raised in the
literaturereview, making it quite clearhow these results confirm or change think-
ing about the problem or issue. It is here that the argumentis most developed and
it is here that the claim to having found something of interest is laid. It can take
about 20 per cent of the report.
Concfusion (and recommendations) The conclusions should already be clear. They
can be restated, but doing only that makes a dull ending. This is a good place to
speculate on the significance of these conclusions- what are their implications,
whether for views of the field, for research methods, or for policy and practice?
The conclusion is a good place to connect your research, which will usually be
small in scale, with far larger issues.Yet the conclusiondoesnot need to be much
longer than 5 per cent of the whole.
References Find out if there is a style that you are expected to use, or choose a
referencing style, and then stick rigorously to it.
Appendices Appendices are not always counted within the length limit. They
are an ideal place to put research instruments, some of the data and a limited
number of other documents of importance. Number them as Appendix 3.1, 4.2
etc. (the first appendix for Chapter 3, which might be an interviewschedule,and
the second appendix for Chapter 4, which might be a particularly interesting
transcript).That said,keep appendicesto the minimum. Toomany appendicescan
be taken as a signthat the writer lacks the power to selectwhat is really important.

omit references.
In response, you may be told about factual errors. Also, this report gives
participantsthe opportunityto confirm that the findings are, at face value, a valid
representationof what they believe and understand, which is a helpful confirma-
tion of the study's validity. Alternatively, where participants have reservations,
these should cause re-thinking (Hammersleyand Atkinson, 1995).However,this
is not saying that participants should have the last word on what your results
mean. The point is that their interpretations should lead you to consider your
interpretation.Once your study is completed,you might wish to distributea final
report to participants.
Practitioner journak
Theseare good channelsof dissemination.They tend to appearfrequently,to have
a need for interesting copy, and to prefer short, clear, practice-focused pieces. If
an articlecan be linkedwith a forthcominganniversaryof interest (50 years of the
NHS, for example),ithas greater appealonthe grounds of topicality. The cardinal
rule for gettingpublishedin thesejournals is to find out what thejournal looks for
in terms of style,typical length and referencingconventions.The secondrule is to
write with thejournal's readershipin mind. This is not only a matter of style,but
one of messages too: readers are likely to be preoccupied with the 'so what?'
question, and this must not only be answered (preferably in the first paragraph)
but also answered in terms of the practical implications of your research. Editors
of these journals can be swayed where there are good, human-interest pictures
accompanyingthe text.
Academic journals
Writing the report 183
The cynic might wonder whether this counts as dissemination,since the number
of people getting beyond the abstract of most articles is surprisingly low (and
estimates of fewer than ten people have been made for some fields andjournals).
However, publication in these journals is a prerequisite for consideration for an
academic post and can be helpful in applying to do research or when going for
other sorts of employment.Mostjournals containguidelinesonwhat they publish
and the preferred style.
Editors of academicjournals will send your articleout to expertsin the field for
refereeing. Ideally, the referees will not know who wrote the article, so they
should judge it on its merits alone. Good referees provide helpful suggestions
for improving your article and do so quickly. Others take months and provide
vague, unhelpful and sometimeshurtful comments.Whether the process is fastor
slow, helpful or hurtful, it is somethingthat happens to all of us in academic life.
Thepoints to heed here arethat it may take a longtime to get your paper accepted;
acceptance may only happen after one, two, or even three revisions; and so your
paper may not be published until two years after you first submittedit.
If your paper is rejected, take account of the reasons, revise with the help of
friends, colleagues and supervisors,and try anotherjournal.
Do not make the mistake of assuming that readers will instantly see the
significance of your work. Unless that is clear, from the beginning and in the
abstract, it is unlikely that most readers of the journal will read your article.For
that reason, editors may be reluctant to consider it for publication. This is espe-
cially true of case studies. If, for example, you have done research on midwifery
in Hong Kong, editors of an internationaljournal will want you to be clear about
the implications of the studyfor an international readerslzip.If no implications
are drawn out, or if they relate to Hong Kong alone, then publicationis unlikely.
Conferencepresentation
It isusually easy to get aproposal acceptedto presentyourwork, usually orallyor
in a poster presentation. With oral presentations, performance skills are very
important. The main advice is never read a paper, although it is good practice to
give participants a paper to take away, or to direct them to the web site where it
may be found. Instead, make up six or so overheadprojector slides,each contain-
ing about six points. Talk to those points, using presentationalskills to best effect
(Gelb, 1988, is excellent on doing presentations). As always, make sure that the
main implications of the research that are relevant to this audience stand out: in
the old adage, tell them what those points will be, tell them the points, and then
tell them that you've told them. Allow time for questions, always in the
knowledge that you will invariably ad lib to your slides for longerthan you had
expected. If, as happens in the best presentations, there are activitiesthat involve
the audience, recognize that these also take longer than is usually assumed.

184 Interviewingfor social scientists 1s'
Bulletin boards
Sharingyour findingselectronicallyis a good way of getting quickreactionsfrom
international audiences. It is important to keep electronic postings short. One
strategyis to post the mainpoints electronicallyand direct interestedrecipientsto
~ppendix
a web site containingthe completearticle(or invite them to e-mailyou for a copy
11 An Interview Research Checklist
of it).Publishingin this way doesnot yet receive much credit in academiccircles.
. I
A book {
Print runscfor wdemic books are normally in the hundreds, which means that f
$ This checklistsummarizeskey issuesdiscussedin the text which you mayneed to
books will not reach large audiences, seldommake your reputation, and certainly think about in your research. However,we have consistentlyemphasized that the
will not make you rich. They are, though, a good way for would-be academicsto
{ research design has to be fit for the purpose, and that research practice is about
startbuilding a reputation.
making the best choices in situations that are not ideal. The implication of this
There is an art to identifyingpublisherswho will be interestedin yourwork and
insistence is that it will be quite reasonable to give a 'No' answer to some of the
in convincingthem of the value of your book proposal (and it is far better to get a
questionswe have raised.
proposal accepted and then write than to send a completed typescript to a
However, since a research report to academic audiences will need to give an
publisher). Most publishers will supply guidelines on making a proposal. They
account of your choice of methods and the way you carried them out, you might
will be looking for an idea that has the edge over other books in the field and that
1 wish to say something in your report about the questions to which you have
will have reasonable sales. Many academic book proposals have very poor sales
answered 'No'.
prospects and will be rejected on that ground alone. Poor sales prospects mean
1 Some questions you will repeatedly revisit, for we have argued that research
that publishersare seldominterestedin publishing theses. One of us has written a
2 'stages' overlap and that certainissueshave to be repeatedlyaddressed during the
book based on a PhD thesis, but substantialpresentational changes were made to
i work. We saw no better way of asking these questions than in this form, but the
the thesis to produce a marketablebook.
1 form itself does tend to suggestthat research is a 'when you've donethat, do this'
process, which is not what we have been trying to say.
!
Conclusion
i
It would be nice to think that this book is completed as these words are written.
Nice, and unrealistic. Before this draft makes it to print, there will be revisions,
and revisionsof revisions,and the referenceswill need to be got into order,which
is always a long and tiresome task. That work will continue alongside our on-
going work on our own interview-basedresearch projects. For Peter, that means
coding up interview data. In some ways a tedious activity, in this case there is
Y
added spice to it, since the two researchers involved are quite independently
devisingindexingsystemsand codingup the data in order to explore the multiple
ways in which the same archive can be read. For Hilary, it means setting up
interviews for the Carers' Act project, refining the research design as the first
couple are done, and beginning to see themes in the data and lines of analysis.
These fluid, overlappingprocesses neatly illustrate the nature of interview-based
research in the qualitativetradition.
B
I8
4