CHECKLIST FOR COMPLETING A GRADUATE-LEVEL SCIENTIFIC RESEARCH PROJECT (A THESIS OR DISSERTATION)

If you’d like a Word version of this exact post, you can find it below from July 2023!

Graduate Research Project Checklist Bajcz A July 2023 Download

Preamble:

Who made this:

Dr. Alex Bajcz, Quantitative Ecologist for the Minnesota Aquatic Invasive Species Research Center (MAISRC) at the University of Minnesota.

Why Alex made this:

Much about how to be a “good graduate student” goes unspoken and assumed. This is particularly true for the scientific research project at the heart of many graduate degrees. A checklist is a valuable tool for ensuring one has “done all that’s expected.” Alex was dismayed to find that such a checklist didn’t already exist when he Googled for one, so he wrote this one. This document attempts to:

Speak expectations and recommendations that might otherwise go unspoken.
Ensure important conversations happen early and often in one’s graduate experience, and that common research missteps can be avoided.
To stress the many steps that ought to occur before any (or most) data are collected!
Provide clarity during what can feel like a murky process as well as encouragement to and guidance on when and how to get ahead (or catch up).
Give mentees and their mentors a tool for strengthening and orienting their relationship (one of the most critical in a mentee’s scientific career!).
Encourage graduate researchers to learn and adopt best practices in study design, data collection and analysis, and communication of scientific results to a variety of audiences.

How to view this document:

As only a relatively complete set of suggestions. Alex based this checklist only on the lived experiences of those close to him. Some crucial elements could be missing, and only very few items on this list should be viewed as “mandatory” in all contexts.
As an adaptable template. Slot in new elements and take out others as you see fit and as best suits your needs, preferences, and discipline.
As general, to be tailored to you by you and your mentor. This document was designed to be as general as possible. Make sure you tailor it to the specific norms and requirements of your department, school, discipline, degree track, timeline, etc.
As divergent. While, generally speaking, Phase 1 items will occur earlier than Phase 3 items, and earlier Phase 1 items will occur before later Phase 1 items, graduate research is often non-linear. Things can, will, and maybe should happen “out of order” sometimes.
As iterative. Graduate research, like all research, is iterative. One should expect to perform some items on this list many times during a research project. View each checkbox as one that may need to be rechecked several times (e.g., once per chapter).
As only one part of a comprehensive guide to graduate school. Graduate life involves much more than just research. Coursework, work-life balance, teaching, etc. are also important but outside the scope of this checklist. However, feel free to add them yourself.

PHASE 1 – STUDY DESIGN

What: Ensuring you are relatively certain about what your project needs to produce before trying to produce those endpoints. Also, preparing yourself and your project team to meet your project’s demands.

When: Ideally, all items here would be complete before gathering any non-pilot data.

☐ Schedule regular meetings (weekly, monthly, etc.) with your mentor, project team, and/or committee for discussing progress, uncertainties, possibilities, challenges, next steps, etc., even if you won’t hold every meeting you schedule. Reserve the time even if you don’t always fill it. Take notes.
☐ Establish with your mentor what a rewarding mentor-mentee relationship looks like for you both. Be specific about your needs and expectations, but also recognize what your mentor can’t or won’t do.
☐ Outline your post-graduation career goals. Work with your mentor to have your proposed project enhance your future employment opportunities. Seek committee members who’ll also enhance them.
☐ Learn to access peer-reviewed literature at your institution and to read the literature to extract relevant information and data efficiently and accurately. Work with your mentor to develop a habit of consuming literature regularly to stay up to date on developments in your research area.
☐ Meet research support personnel at your institution (e.g., your science librarian, the writing center, communications personnel, etc.) to understand and begin using their services.
☐ Identify a peer group (e.g., lab mates, office mates, your program cohort, etc.) with whom you can exchange ideas, especially at preliminary stages and when your mentors are unavailable.
☐ Adopt a reference management platform (e.g., Zotero, Mendeley, EndNote, etc.)¹. Track literature you’ve found and read; share that literature (along with annotations) with your project team.
☐ Browse the relevant literature to familiarize yourself with your study system/subject(s) and with the methods commonly used in projects like yours.
☐ Reflect on what makes you curious about the project. Share your reflection with your mentor and work with them to ensure your project aligns as much as possible with your curiosity.
☐ Articulate your project’s question(s)² (new, insightful answers will be your project’s primary products). Ensure they satisfy requirements for any funding sources you have.
☐ Articulate your project’s hypotheses³, (at least) 1 per question. Ensure these are informed (i.e., have sufficient hard-data⁴and logical support).
☐ Identify knowledge/skill gaps in your project team. Identify potential collaborators, committee members, and mentors who could fill these or seek development opportunities to fill them yourself.
☐ Design your project’s test⁵. Articulate your variables, how meaningful variation in them will be achieved, by roughly how much they should vary, and when and how they should be measured.
☐ Visit your lab/study site(s) early. Document conditions, gradients, challenges, etc.
☐ Develop a supplies list for your project test. Assess existing resources; secure any additional needed.
☐ Seek out funding opportunities to finance your project’s test.
☐Seekout extra help, if needed, to set up your test or to assist you with data collection.
☐ Seek out sources of bias⁶ and confounding factors⁷ in one’s test and attempt to eliminate or mitigate them. Seek outside perspectives from a trusted mentor or data scientist during this process⁸.
☐ Consider ethical implications of your project. Seek regulatory approval (e.g., IRB) as appropriate.
☐ Articulate your project’s predictions⁹ with respect to your project’s test. Then, draw your project’s predictive graphs (i.e., your predictions in visual form). These should clearly include: What data are on each axis (including units!); how those data are to be represented (e.g., as bars, boxplots, points, lines, etc.); and the general patterns/trends expected.
☐ Determine which statistical tools you’ll use to assess the patterns/trends expected in your predictive graphs. Acquire training or mentorship in statistics and statistical software as needed.
☐ Consider possible applied/practical applications of your research. Consult relevant stakeholders to gauge interest and seek feedback.

PHASE 2 – DATA COLLECTION

What: Ensuring you are gathering as much reliable data as possible and documenting your process enough to evaluate, share, and defend it.

When: As lab/field work is being set up and conducted.

☐ Sketch out a general project timeline (e.g., a Gantt Chart). Include when items should be started and completed by and by whom. Set ambitious goals but build in time for rest, delays, etc.
☐ Establish Standard Operating Procedures (SOPs)¹⁰ for your data collection processes. Ensure these are digestible to everyone who may reference them, including potential reviewers.
☐ Establish a data recording system (e.g., rubrics, lab notebook pages, data sheets, a tablet with Spreadsheet software, audio/video recordings, etc.) that is fast and efficient to use under operating conditions. Strive for as little effort as possible to be needed to transfer data to master data files.
☐ Establish a single project folder. This will usually include subfolders for raw and cleaned data, text, presentation, graph, and code files, among others. Use a backup system (e.g., Google Drive, DropBox, external hard drive, etc.). Ensure your raw data files are secure in multiple locations and stay raw¹¹.
☐ Practice your own protocols and train helpers on them in both low- and high-stakes settings.
☐ Finalize your project test. Simplify the design wherever possible.
☐ Set up your project’s test.
☐ Effect (or observe) the intended variation in your explanatory variables.
☐ Record your dependent variable data. Be sure to note anything unexpected, even if it seems trivial. Take pictures and write detailed notes during your data collection about your data collection. Don’t assume you will remember anything pertinent later—document it!
☐ Have a trusted mentor or data scientist observe your data collection and provide feedback.
☐ Once your “main data collection process” is operating smoothly, consider whether you have the time, resources, interest, etc. to collect any “bonus data” for use in alternative/supplemental analyses.
☐ Draft the Introduction and Methods sections of your project paper(s).
☐ Seek out opportunities to “soft-present” your work (outreach events, press interviews, seminar series, etc.). Incorporate feedback from these opportunities into your study design if possible.

PHASE 3 – DATA ANALYSIS

What: Extracting as much new knowledge from the data you have collected as possible while respecting the limitations and assumptions of the tools and techniques you’ve used.

When: After the bulk of the field/lab data have been collected, and following finalization of methods, but ideally before data collection has ceased.

☐ Explore¹² your raw data. Generate summary statistics, exploratory graphs, etc. Look for data entry errors, correlations between explanatory variables, etc. Seek out unexpected data or patterns. Share your explorations with your mentor. Revise your planned analytical approach.
☐ Assess whether “eleventh hour” data are needed to inform future interpretations, bolster one’s analyses, explain unexpected patterns, or evaluate alternative hypotheses.
☐ Wrangle your data into master data sheets specific to each analysis, with the appropriate replicable units¹³ as rows and explanatory and dependent data for those units as columns. Document this wrangling process in a text or annotated code file.
☐ Complete “project takedown.”
☐ Conduct your statistical analyses and render your results graphs/tables, etc. Document the analyses in a text or annotated code file so it’s intelligible to everyone who may later reference it.
☐ Share your results with a trusted mentor or analyst before drawing firm conclusions in order to receive feedback. Revise your analytical approach.
☐ Draft the Results section of your project paper(s).
☐ Prepare data, script, SOP, graph, etc. files for submission to a data repository.
☐ Finish compiling comparable hard data from the literature (i.e., similar system/organism, similar aims, similar methodology, similar patterns, etc.) for contextualizing your results in the Discussion.
☐ Seek out grants, fellowships, or scholarships to support you as you focus on writing your project paper(s) as well as on publishing and graduating.

PHASE 4 – COMMUNICATION OF RESULTS

What: Sharing the knowledge you have gleaned from your analyses with your peers, the scientific community, stakeholders, and the general public so it can be acted upon and so all may benefit from it.

When: As outlined here, begins after all statistical analyses have been completed.

☐ Draw conclusions from your results in light of alternative hypotheses and your project’s limitations. Craft logical and thorough defenses supported by hard data (yours and others). Be appropriately conservative—“proof” is unattainable from just one study, and p values can be (non)significant in error.
☐ Identify a target journal for your final paper(s). Examine its aims and scope and formatting guidelines and ensure your paper conforms to its expectations and requirements as you write.
☐ Draft the Discussion and Abstract sections of your project paper(s).
☐ Provide your mentor with a (near-)complete draft of your project paper with at least 4 weeks lead time before the next deadline (2 weeks for their providing feedback + 2 weeks for your revisions).
☐ Provide committee/co-authors with complete revised draft of project paper with at least 4 weeks lead time before the next deadline (2 weeks for their providing feedback + 2 weeks for your revisions).
☐ Prepare a single master slide deck that contains a complete “narrative” of your project¹⁴. Consider, as you go, how you would abridge it for shorter time slots (e.g., 30, 15, and 5 minutes) and also how you would adapt it for different audiences. Prioritize AV elements over slide text—show, don’t tell.
☐ Consider applying to present at an upcoming conference. Assess if there is financial support you can access or apply for to help facilitate your attendance, travel, or accommodations.
☐ Book a date and venue for your defense/final presentation.
☐ Share your graphs, tables, poster, slideshow, etc. with a trusted mentor, communications expert, or peer group for feedback on clarity, polish, accessibility, audience-appropriateness, etc.
☐ Defend your thesis/dissertation. Revise your final product in response to committee feedback. Submit your final thesis/dissertation to your graduate school, as per its guidelines.
☐ Submit your final project paper to your target journal.
☐ Deliver on any applied/practical applications of your work for relevant stakeholders.

Footnotes:

Check with colleagues first to ensure you adopt shared tools whenever possible.
A scientific question should be system–specific, informed by past work, important to answer, and defensibly answerable. Avoid “yes-no” questions. Questions should somehow concern observable and measurable properties.
A hypothesis is an explanation for how/why something is the way it is. Often, it is a cause of one or more effects.
Hard data are pulled almost exclusively from the Results sections of papers and are not claims or interpretations.
A test is any means of observing how variation in one (or more) explanatory variables affects variation in one (or more) dependent variables. Common types include observational studies, experiments, surveys, and meta-analyses.
Bias occurs when factors or methodological choices that tend to systematically (dis)favor one’s hypotheses go unaccounted for, either intentionally or not.
Confounding factors are those other than your explanatory variables capable of affecting one’s dependent variables.
The ultimate goal is to ensure your data are representative for the population you hope to eventually generalize about.
Predictions are the results you expect to observe in your test if your hypotheses are correct. Good hypotheses should imply at least general predictions, but more specific predictions may emerge once a project test is designed.
Step-by-step “recipes” for how to collect data objectively, consistently, and accurately, to be followed by all on the team.
Raw data files never receive any changes, even corrections. Use additional files for data that have been cleaned, corrected, wrangled, reduced, or reshaped for analysis.
Data exploration is separate from hypothesis testing and model formation. It is about familiarizing yourself with your data and their characteristics and looking for errant or unexpected values and patterns, i.e., anything that could change your analytical approach. Do not conduct statistical tests related to your hypotheses during exploration.
Replicable units are members of the population you wish to generalize about or observations of them. They may be fully independent observations, aggregates of pseudo-replicates, or pseudo-replicated observations if you’re accounting for pseudoreplication in your analyses.
Aim for between 40 and 60 minutes. Aim for 1 minute per slide (2 minutes for graphs!) and for a relatively even number of slides from each of the four main sections of your paper(s). As you cut for shorter durations, cut “whole, lesser stories,” leaving “whole, richer stories.” Don’t include extraneous details. Presentations are primarily for communicating stories, not showing how smart you are or how much work you did!