Figure 7.5An example of median-smoothing – the

1 9 AUTHORING AP H D
Overall, consistently developing and using stem-and-leaf,
box-and-whisker and data smoothing will take very little extra time or trouble for someone who already has a mastery of elementary data-analysis techniques. But you can make very substantial gains in terms of clarifying your own thinking about what your numbers show, especially if you also follow the guidelines set out above for all your tables and charts. Considering carefully what readers need to know can greatly influence and inform the rigour and usefulness of your own thinking. It leads you away from living with complex-seeming numbers that give the illusion of professional authenticity, and pushes you instead towards a genuinely synoptic and insightful analysis of what your data show. Properly understanding data means reducing them to essentials, being able to separate the wheat from the chaff, yet without losing sight of the important details.
Using diagrams and images
Thinking is not just the application of pure concepts arising from a previous verbalization, it is also the entertaining of diagrammatic representation.
Umberto Eco
9
Some attention points are not in the least numerical, but simply visual. They work by using the two-dimensional space of a blank page to permit more complex orderings and representations of relationships than can be easily accomplished by text descriptions. Text is essentially linear. It arranges things in one dominant sequence, but a purely literary explanation is liable to get vaguer and harder for readers to follow as the patterns being described get more complex or convoluted. Here a good diagram can provide an invaluable spine for readers understanding,
allowing them to form a core mental image of how concepts,
events, causes or institutions interact, which can then be fine- tuned and elaborated by your textual account. Readers pay special attention to diagrams of this kind, and they expect them to do a useful job of work. If they do not, if they seem redundant or dispensable elements in your explanation, then readers will be

HANDLING ATTENTION POINTS disappointed. If they put effort into understanding a diagram that turns out to shed no extra illumination, or perhaps is more confusing and less clear than the text which accompanies it, then readers may feel resentful. Such an attention point only detracts from the overall impact of your argument.
So diagrams need to be very carefully designed and implemented. As author you need to ask all the time what readers need to know, what value-added the diagram (or a set of diagrams) gives, and how comprehensible and accessible it will be for them. In short you need to manage readers expectations in a very active way. Some key rules include the following:
◆
Always design diagrams using proper packages appropriate for this task, of which the best known are Microsoft’s
Powerpoint or Lotus’s Freelance Graphics. Simpler illustrations may also be feasible to do in Word or Wordperfect.
Hand-drawn or pencil-and-paper diagrams are no longer acceptable in PhD theses. Nor is it a good idea to use nonspecialist software (like programming languages or general PC drawing packages) to try and produce professional-looking finished designs. If you will need a large number of diagrams, or you will be giving conference or seminar presentations, it is worthwhile going on a proper training course to learn your chosen presentations package thoroughly. It can sometimes be far more time-consuming in the end to try and take a shortcut by hacking-and-seeing or relying on the packages online tutorials rather than attending a few classes.
◆
When designing diagrams always follow well-known conventions for constructing organization charts,
algorithms, or flow diagrams. Conventions speedup your communication with readers, because they can recognize more quickly what is being shown and what you are trying to do, without their having to rely on the accompanying main text. They also impose a discipline on you, preventing you from lurching off into graphical idiosyncrasies which will be impenetrable to readers. For instance, flowcharts outline a sequence of operations or provide a picture of how a set of cases breaks down into several subsets. In the social sciences most authors use a simplified format which focuses

on processes shown in square boxes connected by horizontal or vertical arrows. There is also a more
‘engineering’ style of flowchart, which uses square boxes to indicate processes, diamond boxes to indicate decision points, and rounded shapes to indicate both the start points of a set of operations and outcomes.
◆
Diagrams should be consistently designed across your thesis,
and not vary widely in their appearance. They should look as simple and uncluttered as possible. If you are using boxes try and give them a standard shape so far as possible. Do not adjust box sizes erratically simply to accommodate different- length labels. Try and align boxes on the page using simple row and column (or grid) patterns, rather than spacing them about erratically. Avoid using more types of shapes than are strictly necessary. Different shapes (for instance, rectangles or square versus circles or ovals) should always signal to readers different types of things being diagrammed.
◆
Connect boxes up in diagrams using the minimum number of straight vertical and horizontal lines or arrows. Try to keep kinked or crossover lines or arrows to an absolute minimum. Always avoid diagonal lines or arrows wherever possible, because their slopes usually have different angles.
Diagrams can easily look a bit messy and unprofessional when diagonals run at different angles. Make clear the status of any lines or arrows used to link boxes with a proper key and completely clear labels. Readers should never be left in any doubt at all about what is being shown by connecting lines or arrows. Aline without arrows is generally more difficult for readers to interpret, because it has no directionality in it. Use this device only to indicate cases where two-way flows or linkages of exactly the same kind occur. Non-directional lines are also feasible in those types of diagrams where normal conventions give the line a reasonably clear meaning, as in organization charts or ‘organograms’ as they are also called. Arrowed lines are generally more helpful and informative for readers so long as they know very clearly what an arrow linking A to B
stands for. Does it mean that A causes B? Or that A

Download 2.39 Mb.

Share with your friends: