Article authored by Amanda Neilen
A central part of the thesis is the results section. Here we discuss how higher degree research (HDR) candidates (more traditionally masters and PhD students) in Australia can succeed in results analyses and writing.
Clear and concise results writing
A results section needs to contain points based on; their data-set, their contribution to a relevant scientific theory and at least one novel finding. Furthermore, these points must join to make a coherent story.
This pyramid can be used to derive a clear and concise results section.
Remain focused on presenting your data-set. That is the data you have collected using the methods described. The first section of the results often is the background data. It sets the scene, and is often shown as a table accompanied with a short paragraph. For background field or experimental conditions these are often described as their mean and sd within a sentence, without using a table.
The key to good data presentation is viable, testable hypotheses. The next points in your results section will show how you data is anchored in scientific theories in your chosen field. These theories were mentioned in your introduction. Present your data from descriptive or inferential test in tables and figures to show support or otherwise for these theories. Stay focused on matching your data analyses with your research questions and hypothesis.
Finally, make sure your results show something novel. At least one point should be related to the novelty of your data within the context of the scientific theories you explored. Did you find something new? The most important piece of information (data) will often by shown as a figure. When presenting complex relationships or numerous variables, a good plot, chart or map can make all the difference.
Length and style
In Australia, postgraduate research students in scientific disciplines have pressures to incorporate published manuscripts into the thesis as a series of standalone chapters, with a connecting introduction and conclusion. This affects how the result section is written. Take the time to work out the best option for you, your research and your collaborators. You may decide to write one long results chapter or a short result sections within each chapter or manuscript of your thesis. Check your university requirements.
To write a results section aimed for publication in a manuscript, make sure you check the requirements for your specific journal online. Click here for an example. A tip from Elsevier for writing a manuscript results section is to “Be sparing in the use of tables and ensure that the data presented in tables do not duplicate results described elsewhere in the article.”
If you are sending your results off for peer-review, be prepared for outside feedback and criticism. This differs to the comfort of your direct supervisors. However, this process instills reflection, discipline in research conduct and critical thought. Keeping in mind that the reviewers are human, and sometimes show their biases.
Draw a mud map of your expected results section
Keep length and style requirements in mind, it is time to establish your plan. Your plan should match your research questions and hypotheses, and show were you would put in specific figures, tables and analyses that may be used test those hypotheses.
Start early your mud map early. The time most suitable is when you have selected a topic area, identified gaps in the literature, established aims and hypotheses. Once you have done these part, before going any further, scribble out your result section. Try a few ideas out. Imagining tables describing your data, what types of figures you would include, and what descriptive and inferential statistics you would complete.
This planning stage is dynamic. Think of your plan as a living document. Therefore, you are expect to update your ideas regularly. For example, when* you come across methodological limitations preventing some data collection.
*somehow it feels like we all have this problem.
Use R to create a dummy data-set that would give you significant results and non-significant results. Create plots for your ‘ideal’ figures. Start considering what you would do if the latter occurs, and your results were not significant? What if the explanatory variables were only weakly related? What if outliers and missing values are scattered throughout your dataset? Perhaps, a second research question is needed? Meet with your supervisors, and get you contingency plan in place early.
Creating this plan during the early stages of your thesis ensures you are collecting the correct data for the questions you are asking. This extra work early on, will save you time later and keeps your project focused.
Tip: Finished collect data? Start by writing down what you know now, as a result of your research.
Time management to overcome struggling with complex statistical procedure
- Keep focused with a mud map. Set small, achievable goals and stick to them. If you diverge, it is time to re-think your plan map. Sessions on the computer with R or SPSS open, and best distinguished from ‘ideas’ time.
- Tap into the resources you have available. Australian universities all offer a range of postgraduate support services often providing statistics workshops, style guides, information sheets and/or statistical consultants.
- Take the time needed to learn statistics thoroughly. There are no short-cuts to acquiring a deep understanding of basic statistics. This means if you are still struggling with the basics, talk to your supervisor about your options for sitting in on a first and second year stats course.
- Know when to outsource. For thorny data which is statistically challenging, the expertise you need may not be available at your university. In these cases, often a number of statistical and numerical experts with different specialties may all be required for consultation.
Email us if you need a thesis consultation today.
By completing these four steps you will not waste time on analyses that diverge to your main research questions. Therefore, keeping your thesis submission on schedule.
Start learning and using your statistical program of choice early in your cadetship. Learning a language such as R will ensure you can perform simple data management tasks during your field or lab work, and powerful analyses when it is time to analyse your data. This is particularly important if you plan to continue in academia after your doctorate.
Learn R specific for your PhD with our online data-driven scientists. Click here.