The final stage of a research project is presenting the results. When you are deciding how to report the findings, you need to make some decisions: should there be a written report, or is some other format better? How long should a report be? What should it include? What style should it use? This chapter will help with those decisions.
If you are a good audience researcher, the audience you should think about now is the audience for the research findings: in other words, the clients. The question here is what is the best way to communicate the findings to the clients? What will they understand most readily?
Any good audience researcher will always consider the audience for the research report, and produce the report in a form that will be the most useful for the audience. The usual types of audience for a report are:
All of these groups have slightly different needs and interests. Managers usually prefer an overview, with concise recommendations and precise numbers. The staff of a media organization usually like a fairly detailed narrative report, covering their particular area of work.
The audience also like narrative reports, but at a more general level. Funders prefer to discover that the organization is doing an excellent job, and is spending its money wisely. But they are very sensitive to being hoodwinked, specially if the report is being written within the media organization itself.
If all these groups are to be satisfied, you may have to produce several different versions of the report.
Survey reports are read by a variety of different people, for different purposes. I find it helpful to divide a report into sections, focusing on a particular type of reader in each section...
People who dont have time to read the whole report (or dont want to put in the effort required) will want a summary. This can be done in three parts:
Information on exactly how the survey was done, for anybody who might want to repeat it. This information is often included in appendixes to the main report. It includes the full questionnaire, the detailed sample design, fieldwork procedures, interviewers instructions, data entry instructions, and recommendations for future research: how to do it better next time. If the research is analysed by computer, a list of all the computer files is also useful.
People who work in a particular department will usually be very interested in what a research study found about their program or department, but may not have much interest in other departments. Sometimes I have prepared special short reports for program-makers with a very strong interest in the research. It may be enough to give them a computer-printed table, and show them how to interpret it.
Though some audience members will be very interested in the research findings, most will have only a casual interest. A short summary a page or two of print, or a few minutes of a radio or TV program is usually enough. An alternative presentation for the general public is an interview with the main researchers; this can be more interesting than a purely factual report.
In the past, most survey results have been presented as written reports often 50 to 1,000 pages long. However, a survey report can take many forms. If its in writing, it can be long, short, or released (like a serial) in several stages. It can be produced as a poster, instead of a bound report. It can be a formal presentation, a course, or a workshop. It can take the form of a radio or TV program, or even a theatrical production.
Which method should you choose? Read on: the answer depends on the time and funds available, but mostly on the audiences for the report. (Perhaps presenting audience research findings needs its own audience research.) Lets consider the possibilities, each in turn.
Surveys cost a lot of money. After spending thousands of dollars on a survey, the people who commissioned it often expect a large detailed report - because those are the only results they will see. Because of this expectation, long reports are written. Perfecting these can take weeks, specially if they have a lot of graphs. The printing and binding can take weeks more. By the time such a report is finished, it may no longer be needed - the information may already be out of date.
The problem with big reports is that theyre so daunting: hundreds of pages, crammed with facts and figures - specially figures. Hardly anybody likes large tables of numbers - but this is what most market researchers produce, most of the time.
A typical would-be reader, when given a fat report, will immediately flick through its pages, and decide that it will take hours to absorb. But not now; he or she is too busy. So she puts it aside, and decides to go through it later. But there is always so much work to do: more reports to read, important decisions to be made, lots of meetings. So very often, the fat report is never read, in full.
And you have wasted your time in writing such a large report. (The only consolation is that, in a few years time, when somebody else comes along to do a survey, a few parts of a large report could be very useful.)
How short is a short report? My suggestion: if a report looks short enough to read in full, as soon as a recipient gets it, then it must be short. The maximum is about half an hours reading time, or 20 pages maximum.
A short report has no space for detailed tables of figures but it can invite readers to consult this additional data - which could be kept in a computer file, and printed out only on demand. Its usually no extra work to produce these appendixes, because these are documents (such as the questionnaire) already created, and used when writing the short report.
If a survey has too many questions for a short report, you can write several short reports. That way, the readers will get the first results more quickly. Each report could be a few pages, covering a few questions in the survey. Distribute these reports several times a week, and (as Ive found) theyll be widely read by your users. Though you may have to issue some corrections or additions, the users will become much more involved with the data.
When all the short reports have been produced, you can combine them into a longer report, for reference.
I dont recommend producing one long report, but sometimes clients or others insist on having one. As long reports take a long time to write, it will be weeks before the clients receive their report. By the time the final report arrives, parts of it may be outdated.
In this situation, its advisable to produce a preliminary report, as soon as you have provisional results. Preliminary reports should not be too detailed; a few hours work is enough. Writing a detailed preliminary report will slow down the production of the main report and readers of the first report are likely not to bother reading a full report.
The simplest way to produce a preliminary report from survey data is to take a copy of the questionnaire, and write percentage results next to each answer choice. For questions with numerical answers, write the average on the questionnaire. Dont bother with open-ended questions in a preliminary report these take too long to analyse. To supplement this annotated questionnaire, you could write a one-page introduction, with basic data about the survey: the method used, the exact area covered, the dates, and the sample size.
When you are deciding whether to write a report of 1 page, 1000 pages, or something in between, here are some points to bear in mind.
Usually, between 5 and 30 pages is fine. This applies to the main section of the report, and doesnt include any appendixes. These dont count unless they make the report so thick that people wont try to read it. The more questions are asked, and the larger the sample size, the longer the report must be (and the less likely it will be read in full).
My general recommendations are:
1. If a survey has a sample less than 500, and less than 20 questions, do a short report (20 pages maximum)
2. If a survey has a sample more than 500, or more than 20 questions, do a series of short reports one for each group of questions, produced at least once a week.
3. If you must do a long report (e.g. because a sponsor insists), precede it with a preliminary report, and follow it up with some other way of communicating the information to the clients, such as a presentation, course, or workshop.
Another report format is the poster or wall-chart: one or more large pieces of paper, covered in graphs, summary tables, and brief comments. This is a good format for presenting research results to an organizations staff - specially if it is put on a wall that they often walk past.
The clearest way to produce wall charts is usually with a number of graphs. Each graph shows the numbers that have produced the shapes on the graph, and also has a short verbal summary. A sheet of A1-size paper (about 60 by 85 cm) can be divided into 8 A4-sized panels, or printed as 8 separate A4 pages, which are later joined together. The 8 panels might include:
Posters are read from a greater distance than printed reports, so the type size needs to be bigger: about 16 to 18 points, compared to 12 points for most printed reports. This means that you dont get as many words on an 8-sheet poster as on an 8-page report so choose the words carefully.
Wall charts look good in colour. If you dont have a colour printer, a graph can be printed in black and white, with hollow bars and areas, and these can be coloured in by hand. Its a good discipline to present the most important results from a survey on the equivalent of 8 sheets of paper - but if the survey had a lot of questions, and you dont want to omit any, you could use even larger paper, or several posters.
Live presentations, usingsoftware such as Powerpoint or Keynote, are becoming very common. If computerized facilities arent available, overhead projectors or flip charts can be used instead. Though a computer presentation looks more advanced, it doesnt provide any information that a hand-drawn chart cannot also do.
A typical presentation lasts from 30 minutes to 1 hour, has 20 to 40 slides or overheads, and is presented by the chief researchers. After the presentation, the audience (usually a group of senior staff - often 10 to 20 people) asks questions and the presenters answer them. Audiences find it more interesting to have several presenters than one single voice.
I find the most effective type of presentation displays graphs and figures. The researchers doesnt read these aloud - the audience can see them perfectly well but instead explain and discuss the findings, engaging in a dialogue with the audience. In these dialogue sessions, large blank sheets of paper should be available, on which one presenter writes any conclusions or requests for further analysis.
One problem with giving presentations to broadcasters and media people is that they are used to a high standard of presentation in programs, so researchers must present findings very well indeed to gain the respect of their audience. So unless you are a very experienced presenter, you should rehearse each presentation (e.g. with other researchers) before it giving it to the real audience. After each rehearsal, you usually find several ways of improving the presentation, to make it clearer and more interesting.
Its unusual for a presentation to completely replace a written report, but often the written report is shorter when a presentation is made. Handing out copies of the slides or overheads is not a substitute for a real report too much is left unexplained. What works best in speech is not so effective in writing.
A problem with a lot of presentations is that the audience - typically the staff of the client organization - are expected to absorb all the survey findings in a short time. When a research project produces a large amount of information, and the staff need more time to understand it, it can be more effective to present the results as a course.
Ive organized several of these courses recently, and have found that by the end of them, the participants know far more about how to use the research findings than do people whove only read (or glanced at) a report, or attended a short presentation.
To be most effective, these courses should not be only lectures, but should have the participants working on short assignments, such as drawing graphs of the findings. The courses can last for anything between a day and a week. The ideal number of participants is between 8 and 20.
A workshop presentation normally lasts between a few hours and a full day. The main difference between a workshop and a course is that a course only presents the results it doesnt consider how to use them. A workshop will usually have decision- makers and managers as participants. The participants (usually 5 to 10 of them) not only receive the results, but also consider how the results can be used in changing the programs.
By this, I dont mean a program that is broadcast, but an audio-taped report that is prepared in the style of a program. A lot of radio producers and presenters are people who are most comfortable to communicate by sound thats why they work in radio. They prefer to gain information by hearing it, not reading it. (Survey interviewers often prefer this format, too.)
If the target audience for a research report is geographically scattered - e.g staff in a radio network - producing a tape of the findings is a good way to bring it to their attention. A good way to produce a professional product (because radio people are reluctant to listen to a non-professional sounding tape) is to use an interview format, and have a well-known presenter interview several of the research team about the findings and their implications.
The researchers need to have a large hand in determining the interview questions, because a presenter not educated in social science wont know what to ask.
However, the audio format is not a good one for presenting exact results, so if an audio presentation is made, a summary report or poster should be circulated with it, giving the numerical findings.
Video presentations are also possible. These can be a mixture of interview and slide presentations, perhaps with some shots of the research interviewing and analysis.
The effort of producing such a program shouldnt be seen as wasteful. If you are dealing with TV people, and need to convince them of the results of the survey, video is the format that they are most comfortable with, and will respond to best. Because figures and graphs can be shown in a video, there may be no need for a supplementary written report.
Video "reports" can be surprisingly effective. I produced one in Papua New Guinea in 1991, when doing an assignment for the broadcasting corporation there. I was told that the managers never bothered to read written reports. As the research was qualitative, using focus groups, we videotaped all the focus group proceedings, and produced a videotape of edited highlights from the groups, showing the most common findings in the research participants own words. This conveyed the survey results very clearly. It would have been better still if wed been able to add still shots of graphs, and written summaries and introductions.
The big problem with video reports is the loss of respondents confidentiality. It may be that somebody who watches the video knows one of the respondents, and that this knowledge may harm the respondent in some way. Unlikely, but possible - e.g. in a survey of an organizations staff. If a video report will be made, respondents need to be told this up front, so they can decide how to react - e.g. refuse to co-operate, to not say some things, or to lie.
"Research theatre" is a new way of presenting the results of audience research. So far, it has been used mainly in developing countries, presenting results of participative research to all the people who were involved. This is often done with puppets. Using the Typicality approach described in the Analysis chapter, a puppet "interviewer" can interview several puppet "audience members," who give answers representing a microcosm of the audience. The puppet interviewer can then discuss the interview with a puppet client, explaining more about how the research was done. Graphs can be used as backdrops, and explained by the same puppets.
This is an effective way of presenting research findings to an audience (perhaps of managers) who think they know more than they really do. They seem to accept being "talked down to" more by puppets than by researchers. If the performance can be made humorous, it seems to be remembered better.
The answer to this question depends on...
If the audience (i.e. the people receiving the report) know a lot about research methods, are very interested in the results, and have plenty of time to study the results, then a full written report is best. This hardly ever happens, so instead of a long report, I recommend a short report (for a short, simple survey) or a series of short reports (for a more complex survey).
If the audience is inexperienced with the results of audience research, and has plenty of time, a course is often best.
If the audience includes a wide variety of people with a wide variety of interest levels (e.g. the staff of a TV station) the poster method often works well.
If the audience is poorly informed, and not particularly interested, a video or puppet-theatre report may be best.
If the research is done for a radio station whose staff prefer to communicate by sound rather than in writing, a spoken presentation or taped report can be best.
If a number of decisions need to be made, and the clients are genuinely looking for research advice but arent sure how to interpret it, a course or workshop is often the best solution.