Properly done, expository writing, such as a grant application, ought to involve an exercise in creative thinking--not in creative writing! Expository writing should never read like a novel or a mystery story. The goal of expository writing is to enlighten the reader by presenting the subject matter in a clear and organized fashion. "Exposition"--which is based on the root word "expose"--is meant to educate and illuminate, not titillate.
Before writing a grant application, first make an outline  so that you can be sure your application follows the agency instructions meticulously and includes the information you are asked to provide. Only after you are completely satisfied with your outline and have checked it for good logical flow should you begin to convert it to prose. To do that well, you need to ensure your writing is accurate, clear, and brief. Be sure to put information only where it belongs. Include appropriately designed and labeled tables, figures, and photographs that support your text. Ensure that data, units, and legends in the figures agree with the terminology used in the application. Be objective: Include relevant references to published work even if you do not agree with the findings.
Accuracy is important to maintain your credibility!
Give the reviewer all the information she/he needs.
Don't overstate your case. In particular, don't call something a fact unless it is a fact.
It is wiser to say "good" or "many" rather than "best" or "most."
Substitute "opinion modifiers" with quantitative information:
Instead of "We use lots of test tubes," write, "We use 500 test tubes per week."
Instead of "most" or "many," write, "68% to 70%."
Instead of writing "recently," give a date.
Use the passive voice sparingly, because it evades the issue of who or what "did/does/will do" the action of the verb.
Write to "express, not to impress." It is important to keep the reader with you all the time, because, although some reviewers will likely work in the same specialty as you, others will not. You must explain exactly what you plan to do and show its importance in the context of the field as a whole. If the reader misunderstands, it is the writer's fault, not the reader's fault. The writer must provide the reader with context and use a logical sequence of presentation--the reader should always be able to understand, and easily perceive, how you got from point A to Point B.
In a grant application, make clear:
Understand that every heading, subheading, and topic sentence is essentially a promise to the reader about what she/he will find in that section, sub-section, or paragraph. Write short paragraphs that begin with informative topic sentences that tell the reader "up-front" what you are going to discuss in that paragraph. Use simple verbs and keep them as close as possible to the subject to which they refer. Avoid complex sentences that may be grammatically correct but difficult to understand. Don't use terminology which may be unfamiliar--and irritating--to a reviewer not in that field.
In expository writing, do not use different words for the same "thing" for literary reasons--you may confuse the reader. Avoid or minimize expressions such as "former," "latter," or "respectively," which tend to make readers backtrack. Try to foresee possible alternative interpretations of what you write that may be different from what you intend to convey. It is also important to avoid grammatical ambiguities.
Eliminate unnecessary words: Modern readers want the maximum information in the minimum number of words. For example, change "We made a decision" to "We decided;" substitute "Now" for "At this point in time." Delete phrases such as, "An interesting example which could be mentioned in this regard is ..." It's a good exercise to go through your manuscript and simply delete any word that does not add anything crucial to the meaning of the sentence.
Use transition words such as "In contrast," that act as a sort of "wake-up" call to tell the reader that you are about to change direction. A reader may nod off, for example, while reading through a long list. But by using phrasing like "In contrast, you should never do..." you will tend to "shock" the reader back to a more alert state.
Minimize use of transition words such as "however," and "therefore." They can be helpful but tend to be overused.
Eliminate useless phrases such as, "Needless to say." If it's "needless to say," then why say it!?
Avoid including information that seems important or interesting but is irrelevant in the context of the current text.
"Last week, a lost little boy was brought into the Emergency Room by a policeman with a broken leg." [Who had the broken leg?]
"Sale, 49, was found strangled with a nylon stocking around her neck and bludgeoned to death by Lexington police December 6 in her home." -- Boston Globe, 30 December 1994. [Was she found by the police or bludgeoned to death by the police?]
"For those of you who have children and don't know it, we have a nursery downstairs." [Don't know that they have children or don't know there is a nursery?]
Excessive Multiple Modifiers
"The computer-based integrated decision support environment ..."
These are words that do not clearly refer back to a particular noun: "We noted that many of the cats were ill and all the dogs had hair loss. This finding ..." [Which finding? OR both taken together?]
"It was decided that..." [Who did the deciding? This passive phrasing avoids assigning the responsibility of the action!]
Interpretation of Words
"I recommend this candidate with no Qualifications whatsoever." [Does he have no bad qualities--or is he "woefully inept."]
(Taken from R. Thornton, Lexicon of Intentionally Ambiguous Recommendations, Meadowbrook Inc., 1988)
Some final bits of advice
Think about style: Avoid both self-adulation and apology. Let the elegance of your data speak for you!
There is a lot of psychology and judgement involved in good writing, so you should think about emphasis and impact with that in mind. For many readers you have only a few seconds to capture their attention, so it's best to put the most important message first. For example, in the phrase, "in the previous manuscript we described three methods to ..." decide whether you want the reader to focus on the fact that it was "in the previous manuscript" or that you "described three methods." If the use of three methods is the more important item, it may make more sense to reverse the phrase to read, "We described three methods in the previous manuscript ..."
Remember that tone and attitude are contagious. Be positive and avoid "pompous" language. Avoid words like "unfortunately," which immediately cause the reader to think that something bad has happened or will happen. Instead of saying that you cannot finish something by June 1, write that you will be able to provide the item by July 30! Don't write "nosocomial infection" when you can just as well write "hospital-acquired infection."
Make your manuscript easy to read. Although you must stay within the imposed page limits, leave sufficient white space to provide the reader with ample psychological and optical relief.
When you have finished writing your document, use spelling and grammar checkers, and also do a thorough human proofread. Give your document to at least three people to review before you write your final draft:
Someone who understands your specific research--to check for accuracy. Someone who understands your field but does NOT know about your specific topic--to check for clarity. Someone who is a good editor.
Someone who understands your specific research--to check for accuracy.
Someone who understands your field but does NOT know about your specific topic--to check for clarity.
Someone who is a good editor.
After you get all the copies back, consider seriously all the readers' suggestions but incorporate only those that you consider appropriate. Do one more final proof read, "polish" the final draft, and send it out for review.
L. Reif-Lehrer, "Following Instructions Is Critical To Success Of A Grant Application," The Scientist, 4 March 1996, p. 15.
G. D. Gopen and J. A. Swan, The Science of Scientific Writing. Am Sci 1990; 550-558.