the really important points .m Favor well-constructed paragraphs over bullets—your written report should not resemble a PowerPoint presentation. Even if you occasion-ally use bullets or other highlights for emphasis, continue to rely on the paragraph as the main unit of presentation. Diagrams, graphs, and tables can be very useful for illustrating, summarizing, and emphasizing information. Use them, but use them sparingly, so that they draw attention to important information. Like headings, their titles should tell a story. All their elements should be labeled‘CoMpletely so, that they, can be understood with little or no reference to the text. Be sure you appropriately round all numerical information; the number of sig-nificant figures should be determined by the least accurate data rather than the number mechanically calculated by your computer. Strive to make the diagrams and graphs as clean as possible, avoiding what Edward Tufte calls "chartjunk," ink on the page that does not tell the reader anything new.39 Convey as much information as you can with as little ink as possible. Also, document your diagrams and graphs with sources as the bot-tom. If you created the diagram or graph yourself, then list "author" as the source. We offer Table 15.5 as a summary of ourme.jor suggestions on communication. Be succinct! Keep the text focused on the logic of your analysis. Relegate tangen-tial points and interesting asides to your files;.or, if you think they might be useful in some way to your analysis, put them in footnotes or appendices. Be careful not to cut corners with jargon; if you use technical terms, use only those that your client will understand. (Sometimes you might purposely include unfamiliar technical terms to prepare your client for debate with others—your task then is to explain the terms in the clearest possible manner.) Try to write crisp text. Start paragraphs with topic sentences. Favor simple over complex sentences. Use the a.ctive rather than passive voice to keep your text lively: "I estimate the cost to be rather than "The cost was estimated to be..." Allow yourself time to edit your own text, especially if you tend to be wordy. • EstaWishing Credibility Until you have established a track record as a reliable analyst, you should expect your Clients to be somewhat nervous about relying on your analysis. After all, they are the ones who will bear the political and career risks from following your advice. Therefore,. if you want your advice to be influential, you must establish the credibility of your analysis. You can enhance the credibility of your analysis in several ways. First, make sure that you cite your sources completely and accurately. Of course, you will find this easier to do if you have kept clear notes on your document and field research. Appropriate citation, as noted in Chapter 14, enables you to give credit to others for their ideas (something you have discovered is important to academics) and to provide a good starting point in the event you or some other analyst has to address this, or a related issue, in the future. Second, flag uncertainties and ambiguities in theories, data, facts, and predictions. You do a great disservice to your client by hiding uncertainty and ambiguity, not just because it is intellettrially'dishonest but also because it May set, your client up to, be blindsided by others In the policy arena who have more sophisticated views. After flag-ging uncertainties and ambiguities, you shotild resolve them to the extent necessary to get on with the analysis. (Perhaps a "balance of evidence" will be the best you can do.) You should always check the implications of your resolution of uncertainty for your recommendations. Where your recommendations are very sensitive to the particular resolution, you should probably report on the implications of making a range of reso-lutions. For example, instead of just working through your assessment of policy alter-. natives under your "base-case" assumptions, you might present assessments under "best-case", and "worst-case" assumptions as well. If your analysis takes the form of a t-b Cosenefit analysis, or uses cost-benefit analysis to monetize efficiency impacts in a multigoal analysis, then you should investigate the sensitivity of your results through Monte Carla techniques, which provide a means of assessing the impact of uncertainty ire pararneteis'ori'.the uncertainty of net benefits through silt-Illations:
đang được dịch, vui lòng đợi..
