Describe tables mainly in terms of their substantive implication. Cite numbers only as much as is necessary to make clear what the table shows, and then state the conclusions the numbers lead you to. The point of presenting data is to test ideas, so the data should be discussed in terms of their implications for the ideas (hypotheses) being tested. Simply citing the numbers is not sufficient. On the other hand, you need to cite enough numbers to guide the reader through the table because most readers –including most professional social scientist-are more or less illiterate when it comes to reading tables.
Strive for simplicity. Try to state your argument and describe your conclusions in terms your ancient grandmother or your cousin the appliance salesman would understand. There is no virtue in obscurity. Obscurity and profundity are not synonyms; obscurity and confusion are, at least in this context. As our brethren in the physical sciences know, truly elegant explanations are almost always simple.
Avoid phrases that add no meaning. For example, instead of “We now investigate what inference we can make as to whether A might be said to have an effect on B,” write “Does A affect B?”
Avoid passive constructions. “It is found that X is related to Y” tells us no more than “X is related to Y.” Avoid “A scale of support for U.S, foreign policy was constructed.” Who constructed it, God? Write “I constructed a scale of support for
Avoid jargon when it does not help. Note that I did not suggest avoiding jargon altogether. Jargon, the technical terms of a particular discipline or craft, has a clear function-economy. Use jargon terms when they enable you to convey a point in a sentence that otherwise would require a paragraph. But if ordinary English work just as well, use it. Unnecessary jargon does not make your writing more professional or more scientific, only more ponderous.
Avoid abbreviations. Abbreviations are unavoidable in tables because often there actually is not much room. But the space you save in the text is not worth the cost of annoying or confusing your readers. Psychologists are particularly guilty on this score, and their papers provide good negative role models (“negative role model” is an example of a jargon term that is efficient). Why say “twenty-seven Rs were male” when you can say “twenty-seven respondents were male” or, better yet, “we have data for twenty-seven men,” and so on.
Do not say “we” when you mean “I”. It is pretentious in a solo-authored paper to say “We constructed a scale.” It invites the retort, “What do you mean ‘we’? Got a mouse in your pocket?” However, it is acceptable to use “we” when you mean yourself and the reader, for example, “As we see in Table 3…”
“Data” is a plural word. Never say “data is” but always “data are.” The fact that semiliterate professors (and writers of computer manuals, for example, Stata)frequently violate this rule does not make the violations any more acceptable.
The terms “association” and “correlation” describe relationships between variables, not particular categories of variables. These terms are never appropriately applied to particular cells or particular rows or columns of a table. Useful phrases for describing association include “there is a positive correlation between A and B.” “when A increases, B tends to increase,” and “when A is high, B tens to be high.” Do not say, “there is a positive correlation of A with high levels of B.” Also do not say “A correlates 82 percent with B”-this is incorrect because it is not what the coefficient means.
출처는 Quantitative data analysis: Doing social research to test ideas (Treiman, 2009)
최근 덧글