Assignment IV for Biostats Course VHM 801 at AVC - Fall Semester 2018

The assignment is worth 15% of the final course mark. Please be aware that by handing in the home assignment you implicitly acknowledge to have read and accepted the instructions for home assignments as described on the VHM 801 homepage.

This assignment focuses on the reporting of statistical analysis in published articles. You are required to select one or several published articles, and answer the questions listed below about the statistical methods reported in the article(s). Your assignment must include a paper (electronic) copy of the article(s) used, including the full reference(s). You are free to select the article(s) as you like, with the following restrictions:

• the publication year should be 2017 or 2018 (and the paper must not have been used in the last VHM 801 course, see detailed list),
  
• you are not allowed to know any of the authors personally (which excludes papers you have authored yourself, and probably excludes most authors from UPEI),
• it is not allowed for several students to work together on the same article(s).

For Questions 1-4 you should focus on one statistical analysis reported in your selected article. The analysis must be relevant for VHM 801 and must be reported in the article with an indication of statistical significance (e.g. by a P-value). By an analysis "relevant for VHM 801" we understand that it has been covered in Sessions 1-12 of the VHM 801 course; note that this e.g. implies that the only type of ANOVA included is a one-way ANOVA as covered in Session 10 of VHM 801. (Note: You are not required to explain multiple or all analyses in the article, just the one analysis you are focusing on. It is allowed to include two analyses, but if your answer includes more than two analyses your mark will be based on your answers for the first two analyses discussed.)

1. Identify clearly (e.g. by a specific reference to a section or a table in the article) the statistical analysis you are focusing on, and explain its research question/objective/purpose (as well as you can from the information given in the article). Describe the study type and the statistical design; make sure to include in your description: the observational/experimental unit, the number of observations, and the variable(s) considered in the analysis.

2. For experimental data, describe how the randomization could have been carried out (as recommended in the Bailar & Mostellar (1988) paper); give enough detail to enable another person to reproduce your method. For observational data, outline instead how a randomized experiment could be carried out to investigate the same research question (assuming there were no restrictions on your access to experimental units etc.). For survey data, describe how the sampling was carried out.
   Additionally, for all study types, describe the population the statistical inference (in the article) is intended to be valid for.

3. State the statistical model and procedure used to analyze the data, as well as the conclusions drawn from the analysis. Explain any parts of the statistical procedure not covered by VHM 801 (Sessions 1-12); if needed, ask the instructor for guidance. If major parts of the statistical procedure are not relevant for VHM 801 (as defined above), you should choose another analysis to focus on. Review briefly the assumptions behind the analysis.

4. Give one (critical) question you could ask about the adequacy or validity of the statistical analysis, either as a reviewer of the manuscript or as a member of the audience of an oral presentation of the study. For example, your question could be related to whether one of the key assumptions for the analysis was met. Describe your rationale behind the question. It is important that your question is "reasonable" so you should try to base your question on the information given in the article. If the answer to your question is given directly in the article or can be obtained from the information provided, you should choose another question. If a partial answer to the question is given or can be obtained, you should explain the partial answer and what additional information you are requesting.

5. Give an example of a statistical analysis that you think is not reported according to the guidelines outlined in Session 11 of VHM 801. You may base your example on either one of the items 1-15 of the Bailar and Mosteller (1988) paper or one of the items/statements in the SAMPL guidelines. Describe carefully what the problem is, and how it could be addressed by the authors. Explain to the authors why you think the issue you're bringing up is important (and try to avoid being overly fussy or petty in your critique, e.g. by focusing on an issue where the requested information is implicit or understood).

6. Give an example of a statistical analysis that you think most likely is inadequate for the data, or where the conclusion given in the article is inadequate for the actual analysis carried out. Describe carefully what the problem is, and how it could be addressed by the authors.