Lab 2. Writing a scientific paper

Plan for discussion:

1. Purposes of publications (or presentations): announcement, report, complete description, teaching, education of general public, review.

2. Audience: scientists in general, scientists specializing in the field, students, general public, legislature.

3. Structure of a scientific paper: (1) title, (2) abstract, (3) introduction, (4) materials and methods, (5) results, (6) discussion, (7) acknowledgments, (8) references cited, (9) illustrations (tables and figures), (10) appendix. Internal structure of each section. In what portion of the paper most of the references should be located? What is the order of writing these sections? If you write a paper with a co-author (e.g., supervisor), how would you schedule the work? How many ideas should be presented in one paper?

4. Which title is better?
"Analysis of experiments on mating disruption of Douglas-fir tussock moth using bead dispensers" or
"Mating disruption of Douglas-fir tussock moth (Lepidoptera: Lymantriidae) using a sprayable bead formulation of Z-6-Heneicosen-11-one"
What are advantages and disadvantages of these titles: completeness, self-explanatory, position of key-words, unnecessary words.

5. Is it good to start the abstract like this:
"In this paper we analyzed the effect of hunger level and prey density on movement patters of three species of carabid beetles ....."? Why do you think it is good or bad?

Check characteristics of a good abstract:

Announces the resultDescribes the result
Does not mention methodsGives a list of methodsDescribes methods in detail
Does not include numerical resultsIncludes some numerical resultsIncludes all numerical results
Does not include statistical tests (F, t, P)Includes statistical tests
Does not include interpretation of resultsIncludes interpretation of results

How to write an abstract for a review paper? What are results, methods and conclusions in review papers?

6. What is the purpose of introduction? Which of the following should (or should not) be included into the introduction: (1) problem description, (2) review of literature on the problem, (3) review of literature on methods, (4) discussion of results obtained by different authors, (5) models used for analysis, (6) objectives, (7) advantages and limitations of your approach? If some of these items don't fit into the introduction, then where to put them?

7. Style of scientific papers. Passive voice; usage of pronouns "I, we"; scientific jargon, e.g., "parameterize"; abbreviations; simplicity of sentence structure; references; minimum direct quoting; terminology and term definitions.

8. Methods. What is wrong with the following description of methods: "Polynomial regression was used to analyze the effect of temperature on the duration of gypsy moth larval development. The analysis was performed using the JLM software (SAS Institute, 1992). Data was imported into the worksheet in the ASCII format: column 1 was temperature, and column 2 was time of larval development. Temperature was considered an independent variable, X, and time of larval development was considered a dependent variable, Y. To estimate regression coefficients we selected menu items "Analyze/Fit X by Y".

9. After reading "Materials and Methods" your colleague from another university should be able to repeat your experiment. If there is a simple well-known procedure then it is better to avoid its description. For example, if you use commercial software, don't describe your actions which are specific to this software: what keys did you push and where you click your mouse.

10. Results: statistical significance, direct biological interpretation. Indirect interpretations can be discussed in the "Discussion" section. Illustrations are important. What is better a table or a graph: how to make a decision?

11. Discussion. Discussion is the most important portion of the paper. If there is nothing to discuss, then there is no reason to write a paper! What to discuss: (1) tested hypotheses, (2) limitations, (3) similar results of other authors, (4) predictions that follow from your result (some times they require models). Should the discussion include conclusions or recommendations?

12. Citation. It should be always clear where are words and ideas of the author, and where are words borrowed from literature. Danger of unintentional plagiarism. Avoid "chain citation", try to find the original paper. Don't ignore papers written in foreign languages; spend your time (or money) on translation.

13. What may be the purpose of the appendix: (1) to amplify discussion, (2) to suggest practical applications, (3) to describe technical details that are not necessary for understanding the paper but provide useful information for readers who are specialized in the area, (4) to describe methods, (5) to provide raw data.

14. Logic of the paper. Check if it is possible to subdivide the paper into two or more independent papers. Check that only one main idea is followed in all sections of the paper.

Lab assignment

1. Read the article given to you. Suggest a title and write an abstract for this paper.

2. Compare your abstract with the author's abstract; which is better? What problems did you have writing the abstract?

3. Write a short review of this paper. Use the previous discussion (see above) for evaluating the quality of the paper.

How to write a review?

The following questions should be addressed in your review:
  1. Is the problem important?
  2. Was the problem solved in this study (at least to some extent)?
  3. Novelty of results
  4. Logic and composition of the paper.
  5. Can it be partitioned into several papers? If yes, then how to subdivide the material.
  6. Can the paper be shortened? If yes, what to delete.
  7. Is it possible to suggest better methods?
  8. Are collected data sufficient?
  9. Are data homogeneous?
  10. Does statistical analysis answer the question?
  11. Is there any way to improve statistical analysis?
  12. Are results properly described and interpreted?
  13. Is it possible to present table data in a graphic form?
  14. Is there any important information that is not presented (e.g., mean values, sample size, SE)?
  15. Are major conclusions explicit and clear?
  16. Are limitations of this study discussed?
  17. Are mechanisms that cause the observed phenomenon discussed?
General evaluation of the paper:
  1. What audience may be interested in reading this paper?
  2. Should it be published?
  3. Does it require a minor or major revision?
Be very specific in your comments and make suggestions how to improve the paper.

Specific rules for lab reports

  1. Title, author, affiliation.
  2. Abstract: describe main results (50-120 words).
  3. Introduction: describe the problem, and your objectives.
  4. Materials and methods. Describe the source of your data, how the data was obtained. Include 1-2 references on the biology of the species and on its economic importance (if any). Describe the model you use and make references to the author of the model. Describe statistical methods or validation methods. What is the rationale for your methods?
  5. Results. Include illustrations (figures and/or graphs) and describe what do they mean. Make figure captions! Axes, legends should be clear.
  6. Discussion. Discuss the problem in general. What did you prove with your analysis? Discuss limitations.

Alexei Sharov 12/4/98