I Have To Do A Research Project

In undertaking a smooth, achievable research project as a student, you should understand that you NEED one thing:

  • a question which has been well-defined with the mentor, and can be answered from the pre-existing data source

Which of course presupposes two necessary conditions to achieve this

  1. a useful mentor who can guide you through the process
  2. an available, pre-existing data source

Failure to achieve the first places you at risk of posing unanswerable, onerous questions. Failure to achieve the second places you in the situation where you have to devise data structures and collection, as well as undertake the collection itself. If any of you know about Sod's Law, you should not put yourself in the position of finding that the common condition will only be common before and after your research period.

Here is what I do when some forlorn padawan comes and asks to do a research project.

  1. if I have no data, I tell them to go away! Fortunately, I do now have data, so I think up a question based on the data I have, and frame the question for the padawan.
  2. I ask the padawan to find me a retrospective report in he literature addressing this question or something very like it.
  3. Then I ask the padawan to use the bibliography of the article to start searching the area and form the basis of their literature review. As a rule, I think a literature review should have >20 papers, and should be undertaken using a Reference Manager
  4. Following on, I ask the padawan to analyse the paper and list the data elements used in the paper, and the analysis that they will undertake on these data elements. e spend some time discussing and 'fixing' these requirements.
  5. The padawan writes up an ethics application to undertake the research.
  6. While waiting for approval from the Ethics Committee, I construct a spreadsheet from my data source to match the data elements required by the project.
  7. On receipt of ethics approval, I send the spreadsheet to the padawan, and set a time for the first analysis. This also takes some discussion, teaching and iterative review.
  8. When analysis is finished, the padawan writes up their report, and presents it for review.

In all cases there are some truisms that will bite you if you don't comply:

  1. data you must have padawan! Seems really obvious to say, but in all cases it's the hardest thing to actually arrange. In many circumstances, you may have to build the infrastructure that provides the data and that can take year or decades. If you are a student looking to do a research project, spend 2-3 weeks as soon as you know it will have to be done LOOKING FOR DATA SOURCES. In terms of learning how to use statistical tools, confront ethics and undertake analysis, it is easier to start with data and not be bogged down in data collection which sits between research question and research analysis.
  2. edible data to software you must present padawan. Any of the statistical packages tht you might use require you to present data to it in an ordered format. whether the file is an open format such as CSV (comma separated variable), ODS (OpenDocument Spreadsheet) or a proprietary format from the dark side such as XLS/XLSX from Excel, the arrangement of the data inside the file has to be just so! Get i wrong and you will have no analysis, get it right and the speed of analysis will make your head spin.
  3. data you must not massage padawan. Beware the urge! It is unscientific. The analysis you undertake must predetermined otherwise you will waste time looking for significance. Get used to the idea that the answer to your question from your data may be "NO", that's OK, just leave it at that. You asked a question, you got an answer! What more do you want?

Remember this is not a PhD, it's student research project trying to teach you the fundamentals of research and analysis, so that you are equipped for the future.

Unless otherwise stated, the content of this page is licensed under Creative Commons Attribution-NonCommercial-NoDerivs 3.0 License