Steven Holland

Project Report

Please follow these instructions carefully. I recommend printing this page and checking each item before submitting your report.


Your report should follow this outline, with headings for each of these sections, in order:

scientific question to be answered. This should say in words what scientific (not statistical) problem you are trying to solve. Provide the necessary background so that the origin of your hypotheses is clear. Start from the general and work to the specific, ending with clearly stated hypotheses. Keep this section brief but detailed enough that any scientist could understand it. Watch for jargon.

data used to address the question. In a concise statement, report the sample size and variables measured. For any non-obvious variables, briefly state what they measure.

statistical methods used, showing that you verified the assumptions of your analyses. Review your notes and your textbook for the assumptions of your chosen tests.

results of your analyses. Include the necessary figures to make your points, but no unnecessary figures. If you must perform many tests, summarize those results in a table. Where you state results in the text, provide numerical support parenthetically. Emphasize estimates and uncertainty, not p-values.

discussion of your results in terms of the scientific question. Explain what your results mean for the scientific question you posed.

references that you cite. Remember the citation rules: if a statement, concept, data, or figure did not originate with you or is not common knowledge, you must cite a reference.

tables that summarize your important results.

figures that show your important results.

Formatting your report

Your report should be prepared as a single .pdf file, containing a cover page, text, references, tables, and figures. Do not include R code or data files in this pdf.

The cover page should include your name, date, and report title. The title should read like the title of a journal article, not “My 8370 Report”.

The text of your report should be three to six pages long. The title page, reference list, figures, and tables do not count towards this page limit. If your first draft is shorter than three pages, you are likely leaving out important details and discussion. If your first draft is over six pages, summarize the results in tables and edit your manuscript. Eliminate unnecessary verbiage while preserving the important arguments and conclusions.

The text of the report should be double-spaced, in Times New Roman, with a 10-point font and one-inch margins on all sides. This should give you 28 lines per page. Do not use creative formatting to bypass the 3–6 page limit.

Include a reference list, consistently and closely following the style of a journal in your field. This list does not need to be long or exhaustive, but you should cite those previous studies necessary to justify the scientific statements in your paper. You do not need to provide references on the statistical methods unless you use methods not covered in class. Do not cite websites unless specifically instructed to do so; cite published literature instead.

All tables and figures should be referenced parenthetically in the text (e.g., Table 1, Fig. 5); do not use constructions like “Table 1 shows…” or “Figure 5 shows…”. Tables and figures should be cited in numerical order. Include only those tables and figures necessary to support your arguments.

Tables should be formatted as they appear in a journal article; do not wrap your tables in a spreadsheet-like grid. Tables should have titles (e.g., Table 1. Summary of statistical tests).

Figures should follow Tufte’s principles. Figures should have captions that state what is shown (e.g., Figure 1. Scatterplot showing the linear relationship between beetle length and width.) and that explain any graphical aspects not immediately obvious. Each figure and its caption should be on its own page. Although figures should be generated in R, they can be modified in Illustrator or similar apps. Your figures should be crisp and clear. I highly recommend against exporting your R plots as .png or .jpg files, as this will cause them to be blurry. Adding PDFs to a Word document often also makes them blurry. A better workflow is to write in Word, export your completed text document as a PDF, and then add the plot PDFs (in Preview on a Mac or in Adobe Acrobat on any platform).

Proofread your paper. Your report should be free of spelling and grammar errors, and I cannot recommend Grammarly (free) highly enough. It should be well-organized into sections and paragraphs. It should include section headings corresponding to the above structure (e.g., Scientific Question, Statistical Methods, Results, Discussion, References). Paragraphs should have well-constructed topic sentences. Spell out state names; do not use their two-letter postal codes (e.g., GA, OH, NY). A space should always be placed between a value and its units (e.g., 2 m, not 2m). Numbers or text that express a range (e.g., 2–7 g, Ordovician–Devonian) are separated by an en-dash, not a hyphen.

Commands file

All analyses and figures should be generated in R.

You should prepare a single text file with the commands needed to produce your data analyses, including those figures generated from your data with R commands. This file should be edited to include all the commands needed to reproduce your work and only those commands. Remove all errors, mistakes, false starts, help calls, and other extraneous materials from your commands file. When I run your commands file, I should be able to replicate your work without errors.

Your commands file should be well-organized and structured, as we have discussed this semester. Use single blank lines and comments as necessary, but do not overcomment your file. Comments should be used to explain blocks of related commands, such as # Figure 2. Comments can also be used for single lines of code that are confusing or whose intent is unclear (you will likely have few, if any, of these). Do not comment on commands whose purpose is obvious.

If you require any libraries, those commands should all appear at the top of your commands file. Do not embed paths in your code. Use loops only when they are the best solution, and follow the advice in Becoming a Better Coder.

What to turn in

Please submit your report in two emails as follows:

The first email should attach your report as a single pdf file; its subject line should be 8370 report. The report file should be named XXXXReport.pdf, where XXXX is your last name (e.g., hollandReport.pdf). Do not submit a Word or other file type.

The second email should contain your R commands file and all necessary data files as attachments, and the subject line should be 8370 analyses. Your commands file should be named XXXXCommands.R, where XXXX is your last name (e.g., hollandCommands.R). Using a .R suffix like this helps you to identify your commands file; it also enables syntax coloring in many text editors. Your data files should be .txt or .csv files, with names that match those in your R commands file. Do not submit Excel files.

You may submit each of these emails only once; be sure that you are completely satisfied with everything before submitting your work. Check again.


Your work will be graded on five areas, each worth one-fifth of the grade.

Ambition of the project undertaken

Appropriateness of statistical techniques for the problem

Correctness of analyses

Interpretation of results in light of the scientific question

Presentation, including following instructions.

Due date

Your pdf report, commands file, and data file(s) are due on Tuesday, 5 December, at 6:00 PM. Late work will be penalized 10 points per 24-hour period after the time it is due.