Chapter 5 makes the point that a technical writer needs to possess basic research skills, else there will be little to write about. Research may take the form of experimentation with new equipment, but it may also require standard library skills, which the text goes over in great detail.
The chapter begins with a four step plan for conducting research. (It says there are five objectives, but there are only four.)
As recommended in the text, we will take a trip to the school library. While there:
The text explains two book classification systems: the Dewey Decimal System and the Library of Congress System. It is good to be aware of each of them when browsing through libraries. Knowing the system a library uses will lead you to the right shelves for your topic. I do not mean that you need to memorize these systems. You should learn how they work, and how to use them. Libraries typically have charts of their systems handy. It is enough to be familiar with the systems, so that you can use them when needed.
It is also useful to be aware that there are specialized reference works available for many subjects. Subject oriented dictionaries, encyclopedias, manuals, and handbooks exist. The text lists several examples of each.
It is possible to use Internet search engines correctly and still find no useful information about your subject. This is due to the fact that anyone, anywhere, can find a way to post information on the Internet. Just because it is there, does not mean that it is correct or authoritative. It is fine to use the Internet for research, but be a little skeptical of your results. I am reminded of the words of an English teacher, many years ago, who cautioned my class to be mindful of who is saying a thing, and to question why it is being said. (Thank you, Mr. Pawloski.)
Regarding search engines, there are new ones available constantly, just as there are new web sites constantly. Your text lists some good search engines. As you should be aware from my site, I think very highly of Google. Another site that is often helpful is Dogpile. Its name isn't very pretty, but its results are. It runs your search query through Google, Yahoo, MSN, Ask Jeeves, and other search engines.
In addition to search engines, the text provides a list of web sites that may be of use to researchers. While this list is interesting, it is still a list of specific sites that may not apply to your current needs. A more general site that is often helpful is the Internet Public Library. It is particularly useful to me for its access to magazine and newspaper web sites, which are helpful when searching for technology news.
The author devotes several pages to explaining a time honored technique of note taking. Note cards may be used to keep track of the references you find. They are useful when you look up a source for more detail, when you need the required information for a bibliography, and when you are organizing your thoughts for your project. This use of note cards dates back to a time before word processing, which is mentioned earlier in the chapter as an alternative to the card technique.
Whether you use note cards, electronic notes, or just a legal pad, you should follow the author's recommendations about documenting your sources to avoid accidental lapses into plagiarism. She offers five rules to keep you honest on page 172:
The author discusses conducting surveys to gather information. When it is appropriate to do so, a survey can be a valuable tool. It can also be completely worthless if it is badly designed. Consider the example survey questions about cookies on page 181.
Example 2 shows a multiple choice question, asking the user to select "the one cookie brand you most prefer". The question is fine, but the response choices are flawed. Hydrox, Oreo, and Fig Newton are specific cookies. Pepperidge Farm, Almost Home, and Duncan Hines are not. They are product lines, not products. If I like some Pepperidge Farm cookies, and dislike others, how do I answer the question? This is the classic problem of comparing "apples and oranges". The same problem exists with example 3, rank ordering. If the choices are not valid for multiple choice, they are not valid for ordering. If you compare inappropriate items, the data you collect will not mean much.
Example 4, a continuum scale, is a familiar type of question to most of us. It illustrates a classic error. The user is asked to choose "the frequency of your cookie purchases per grocery trip". The choices are: always, often, seldom, or never. It is likely that respondents will have a common understanding of the terms "always" and "never", but it is unlikely that there is a common definition for "seldom" or "often". If people were to discuss these choices in a group, a concensus could be reached, but survey takers need definitions on the survey. A better choice might be to offer the respondents a numeric scale of percentages, and ask "on what percentage of shopping trips do you buy cookies?"
A well constructed survey instrument should be tested, adjusted, and retested until it performs the function it is meant to perform.
Surveys can lead to another data gathering technique, one that is often used by technical writers, the personal interview. A survey may be used to identify the right people to interview about a product, a process, a program, or a problem. There is no substitute for working with an expert about your topic when you are gathering information for a technical document. Be aware that personal interviews are the most time intensive method of gathering information, so use due care in selecting who you interview.