Topics

Structure of a paper

A scientific paper represents a highly stylized form of storytelling in which the story is broken into distinct sections, each with its own function. The most common sequence is title, abstract, material and methods, results, discussion, acknowlegments, and literature cited. When done well, this sequence lends an obvious and logical progression to the story. What's even more important today is that this structure enables the reader to find specific information he or she needs without wading through the entire paper. If you want to know about a detail of recording an electromyogram, you can find it in the material and methods section. If the abstract mentions a relationship between oxygen consumption and heart rate, you can turn directly to the results section to find a graph of that relationship, and the legend of that graph should contain enough information to allow you to interpret what you see even without reading the text.

Each journal specifies the finer details of structure it requires, even down to formatting details such as whether to abbreviate "Figure" as "fig." or whether to put a comma between the author and year in a reference. These details are given in the journal's "Instructions to Authors" document. Though they may seem trivial, these formatting specifications must be followed exactly or the submitted manuscript will be rejected without even a review of the content. The instructions to authors for Physiological Advances and Retreats, the imaginary journal of IB132L, are patterned on those for the real journal Physiological and Biochemical Zoology.

back to Papers

How to present statistics

The general rule is to present just enough information so that a reader familiar with the statistical procedure understands what you did and can assess the validity of your conclusion. This is a lot less information than is typically presented in a student lab report. In some cases a p-value may be all that is needed. In more complex analyses, a statistical table may be appropriate (such as an ANOVA table), but you generally should not include all the output provided by the statistical software.

Regressions: You can include the regression equation in an X-Y scatterplot or in its legend. In the text, back up your statements with some statistics from the regression analysis.

Student lab report (a very good one!):
"We performed a linear regression of peak electrical activity on walking speed and tested the null hypothesis that peak electrical activity is unaffected by speed, meaning that the slope equals one. We found that the slope was significantly greater than one, as shown by the p-value of 0.012, causing us to reject our null hypothesis and conclude that there is a significant relationship. The R2 of 0.91 indicates that most of the variance in peak electrical activity is accounted for by walking speed."

Paper:
"Peak electrical activity varied linearly with walking speed (Figure 3, R2=.91, p=.012)."

Comparing means: For simple t-tests and one-way analyses of variance, the statistics can be included in the text.

Student lab report:
"We tested for differences in heart rate among the control and the two hand exercises with a one-way ANOVA and the results are presented in Table 2. Since p= .13, we accept our null hypothesis that the exercises do not affect heart rate."

Paper:
"The hand exercises did not significantly affect heart rate compared to the resting control (ANOVA, F=2.34, p=.13)."

Discussing "significance": In scientific writing the word "significant" is reserved for differences or relationships that are shown by a statistical test to be larger or stronger than could be accounted for by chance alone. Do not use this word unless you have the statistics to back it up! You may wish instead to discuss whether an effect is important, meaningful, or substantial. Keep in mind that a significant difference may or may not be important or meaningful, depending on the context. back to Papers

Writing style

Choose the right tense. When discussing your methods and results, use the past tense. "Reaction time increased as the temperature of the arm decreased". However, when discussing a fact or relationship that has become established by previous publication, use the present tense. "Nerve conduction velocity increases with temperature (Boyles 1972)." There are exceptions to this general rule, such as referring to statistics or figures in your own work "This difference is significant . . .", "Figure 3 indicates . . .", or specifically describing previous work "Boyles (1972) showed that . . ." Koffman used a 0.2 M glucose solution to . . .".

Use active verbs and active voice. If you look for examples of turgid writing in journals, chances are you'll find the authors employ passive verbs, passive voice, and nominalizations. Clear writing, by contrast, uses subjects and verbs that convey the action of the major character or agent in the sentence. Consider this passage:

The employment of a recently developed model (Franklin 1998) for determination of spinal stability in vivo, within the framework of a prospective study, would perhaps be helpful in clarification of the importance of this potential risk factor in the development of first-time low back pain.

The subject of this sentence, employment, is not really the agent that acts, and the verb, be, does not express an action, but merely indicates that an action exists. Note how many of the nouns represent verbs or adjectives that have been nominalized (converted into nouns): employment, determination, stability, clarification, development. By changing the subjects and verbs to reflect the agents and actions, the sentence becomes:

Researchers can use Franklin's model (1998) to determine whether persons with spinal instability later develop low back pain.

Scientists have traditionally used the passive voice extensively in their writing, but this is changing. Active voice, by making the agent of action clear, is usually more precise and less wordy.

Passive: "It was found that . . ."
Active: "I found . . ."

Passive: "Twice as much force was produced by the muscle . . ."
Active: "The muscle produced twice as much force . . ."

Journal editors typically favor active verbs and active voice because they tend to save space as well as improve clarity. Many leading journals, including Science, allow the use of first person, though its overuse can be monotonous.

Avoid slang, euphemisms, and excess jargon. Consider the needs of your audience when using non-standard terms and phrases. English-language journals are read by scientists all over the world. A slang term specific to a particular culture may be misinterpreted by a reader outside that culture. Scientific writing in academic journals necessarily contains some jargon, but this should be limited to terms known by the typical reader of the journal, not just the specialists in your sub-field. back to Papers

Using the literature

What do we mean when we refer to "the literature"? Some students consider the textbook as the primary source of knowledge. More accurately, it is a tertiary source. The true primary sources are short papers published in scientific journals. Each paper ordinarily reports the results of just one or a few closely related studies. This is the type of paper you will be writing to report your results. A secondary source of scientific knowledge is the review paper. Reviews attempt to summarize, integrate, and evaluate large numbers of primary research papers. Reviews appear in scientific journals, in serials dedicated to reviews (e.g., Physiological Reviews, Annual Review of Physiology), and in edited multi-author books. Finding a review of an area of interest could conceivably save you considerable time. Textbooks designed for course use represent a third level, or tertiary source of information, in which one or a few authors summarize and integrate information on a broad array of topics. Textbooks are usually easier to understand for the non-specialist, but generally offer less documentation (references to the primary literature) and a less authoritative view (in areas outside the author's immediate specialty) than do primary and secondary sources. Keep in mind that recent primary sources offer the most up-to-date information because of the lag time required for reviews and textbooks to appear.

Because the literature is so immense, finding material on a topic of interest can be a challenge. One strategy is to start with your textbook or a single recent paper on your topic and peruse its literature cited section for interesting-looking titles. This will usually lead you deeply into some relevant primary and secondary sources, whose literature cited sections will lead you still further. One drawback is that this approach invariably takes you back in time to older and older sources. Sources published subsequent to or even contemporaneously with your first references are not going to come to light this way.

Another strategy that gives you access to more literature, particularly to recent publications, is to use various bibliographic search tools. Medline and Biosis are two computerized databases that index other publications. If you know the name of an author, the title of a paper, or a key word, you can search these databases for bibliographic citations. These databases are available on computers in the Biosciences library and over the campus network on the web. Web of Science (formerly Science Citation Index) is like looking up cited references in the back of a paper but in the reverse direction. You begin with a paper related to your topic (e.g., Adams and Dolan, 1995. Recent advances in lumbar spinal mechanics. . .), find the paper in the search engine, and perform a "cited reference search" to learn what other papers have cited the Adams and Dolan paper since it was published. Chances are that recent papers will deal with the same subject as the cited paper. Ask the reference librarian or your instructor if you need help learning to use these tools. back to Papers

Citing references

There are several reasons to cite the publications of others. The most important is to give credit to the person(s) whose ideas you used in your research. Another is to make your statements more credible ("if you don't believe me, here's where you can look it up!"). A third reason is to save space in the journal by, for example, replacing a description of a technique with a reference to a description published elsewhere.

You will notice in reading the scientific literature that direct quotations of other publications are almost never used. Instead, you describe the concept or fact in your own words and refer to the source article or book listed in the literature-cited section. If you obtain from another person an idea or observation that is unpublished, you should credit it as a personal communication (e.g., Ghost crabs search for prey on open beaches (Wright, pers. comm.)). It is preferable to cite a published source if one can be found, since the reader can better judge the credibility of a source that is accessible for scrutiny.

There are several common styles used by journals for embedding references in the text (and many small variations on these). In the name and year system (used by Physiological Advances and Retreats), the reference is indicated by enclosing the last name and year of publication in parentheses. The literature cited section then is organized alphabetically by author's last name. If there are two authors, both last names are given in the text citation, but if there are three or more, the first author's name is given followed by "et al.". The author's name is omitted from the parentheses if it was used in the text of the sentence. Here are some examples of this citation style:

The relationship between electrical activity and force is sometimes linear (Erlenmeyer 1984; Frankel and Jones 1992), but more often nonlinear (Wu et al. 1986; Johns 1994; Williams and Takahashi 1999). Schumacker (1994) suggested that . . .

The citations should be placed immediately after the idea or observation to make the connection clear. In this example, if all five citations in the first sentence were listed at the end of the sentence, then it would be unclear which documented linear relationships and which nonlinear.

In the alphabet-number system, each reference in the text is indicated by a number in parentheses. The number refers to a numbered alphabetical list by author name in the literature cited section. In a third style, the citation order system, references are cited by number in the order in which they appear in the text. The literature cited section is ordered accordingly by number. These last two styles have the advantage of saving space in the journal and reducing the clutter of names and years in paragraphs that review an extensive literature. It is awkward for the author to place numerical references in the text until all references have been finalized (though bibliographic software now eases this task considerably). Some readers prefer the name and year system because it allows them to associate the cited concept with a name and year without turning constantly to the literature cited section. In any case, the journal dictates the style of references in its instructions to authors. back to Papers

The publishing process

Choosing a journal. Scientists select a journal to publish in based on the audience they wish to reach and the quality of the research. Journals vary in the degree of specialization (e.g. science, biology, physiology, muscle physiology) and perceived quality. If an author has an exciting result based on convincing evidence, it is to her advantage to publish it in one of the top-ranked journals. Science and Nature are the most prestigious and widely-read scientific journals, but they publish only a tiny fraction of submitted papers and then only in a very abbreviated format. Among more specialized journals, quality ranges widely, from top journals found in most academic libraries and read regularly by most researchers in their field, to journals that are hard-to-find, publish lower-quality papers, and are less widely-read. The availability and readership of journals also varies with price. Journals published by not-for-profit scientific societies tend to be less expensive than those published by commercial publishers. More individuals and libraries subscribe to the less-expensive journals, so readership is correspondingly higher. Scientists generally try to publish their articles in the best journals possible, but even outstanding researchers suffer paper rejections and publish some of their papers in lesser-ranked journals. Science proceeds in small steps as well as big ones, and sometimes those small steps turn out to be very important, even if they're not reported in the top journals.

What happens when you submit a manuscript? The journal's managing editor typically examines the paper to see whether the topic is appropriate for the journal and whether it conforms to the format specified in the instructions to authors. If the manuscript fails either of these tests, it is returned to the author without review. If it passes, the manuscript is assigned to an editor, usually an eminent professor serving without pay. The editor then sends copies of the manuscript to two peer reviewers, other scientists not affiliated with the journal, who are qualified to evaluate the paper. These reviewers write a critique of the paper and recommend acceptance, acceptance conditional on major revision, or rejection. If the two reviewers differ in their recommendations, the editor may request an additional outside review, especially if the editor is not a specialist on the topic of the paper. In the end, the editor decides the fate of the paper. back to Papers

What if my paper is rejected? Sometimes a paper is ruled unacceptable, but the editor encourages the author to revise the paper substantially and resubmit it (see below). In other cases, the rejection is final, and you should either try another journal that may be more receptive or consider how your study can be improved. The reviewers may have identified flaws in your study that can be overcome by further experiments or observations. It is also possible that one or (rarely) both reviewers were biased and recommended rejection for reasons that are invalid. You can appeal an editor's decision by challenging the substance of the reviewers' comments, but this is not likely to succeed.

What if my paper needs revision? The vast majority of published papers were initially rejected (with resubmission encouraged) or accepted pending major revision. The review process is not about just weeding out unacceptable papers, but improving the good ones. Take advantage of the comments of your peers to make your paper even better. Of course, you may find some instances where a comment seems inappropriate, even erroneous. Reviewers are not infallible and editors know this. If you strongly disagree with a criticism or suggested change and it materially affects your message, you may want to clarify the disputed statement and rebut the criticism. In your cover letter accompanying the revised manuscript, you explain how you addressed each of the reviewers' points, including any that you chose not to accept. Again, it is the editor that makes the final decision whether to accept your revised draft or return it yet again.

back to Papers

IB 132 Lecture Website

Department of Integrative Biology

U. C. Berkeley

Send us your comments on the IB132L pages!


Last updated 1/21/04
Copyright © Department of Integrative Biology. All rights reserved 2004