Studying for a PhD: Presenting Your Results from Quantitative and Qualitative Studies

Studying for a PhD: Presenting Your Results from Quantitative and Qualitative Studies
Page content

Presenting Your Results

How you present your results will depend upon the type and quantity of data that you have generated. The purpose of the dissertation results section is to allow others to judge the reliability of them, and to provide an evidence base for the conclusions that you will draw in the rest of your thesis. Therefore, you must present your results in a manner and in sufficient depth to meet this goal.

Can You Ever Present Too Much Data?

The short answer to this question is “Yes!” In quantitative studies, some types of problems, for example, computer modelling may generate too much raw data to include. If you simply include everything, you will lose the significant results amongst the sheer mass of figures. Part of writing up a PhD is to tell a story to the reader; huge amounts of raw data distract from the story being told. Similarly, qualitative studies generate huge amounts of raw data and including everything may hide the significant results or stop the flow of the thesis narrative.

Therefore, part of presenting your results is to do so in such a way as to point the reader to the key points that they need to be made aware of. However, this may involve results included either to provide context or to highlight significant negative results.

In order to provide an audit trail for the reader to verify that you have presented your results fairly and accurately, you may wish to include complete data as an appendix or as an electronic dataset provided on a CD-ROM inside the back cover. As always, you should provide a rationale for why you have summarised your results as you have and the choices you have made for selecting or summarising results to be included in the main text as opposed to in an appendix or electronic dataset.

Presenting Quantitative Results

In general, quantitative studies generate less data than qualitative studies. When datasets are small, it may be possible to present your entire results without summary in some cases.

However, when distributions are involved, simple descriptive statistics can actually add value to a presentation. Thus,

“30 values with a mean of 15, and a maximum of 19 and a minimum of 12” may communicate better than

“17 13 16 14 12 18 15 16 14 13 19 17 14 16 18 12 14 12 13 18 17 14 15 16 14 15 17 12 13 15”

Tables can be a good way of structuring the presentation of data. They can draw attention to key values when they might get lost in a prose description.

Graphs are good at showing trends or relative values, but can lose detail of actual values. For example, they can illustrate the responses to a Likert scale question and show the relative proportions of responses.

However, there are some common pitfalls. Students often seem to feel obliged to “sex up” their data by using unnecessarily complex statistics or software tools. SPSS is not an appropriate tool to calculate a simple mean. If you are going to use it to more fully describe a distribution, in terms of, for example, a mean and a standard deviation, then you need to understand the difference between parametric and non-parametric statistics and be able to show that your data conforms to a normal distribution.

Charts can cause trouble, too. Again it’s good advice to keep it simple. 3-D charts look nice but actually obscure the results. There is evidence to suggest that people find pie charts difficult to read.

And, finally, be careful of percentages. It’s always worth quoting the absolute value as well. 5% of 20 quoted as 4 (5%) or 5% (n=4) is often a very different result from 5% of 20000. Percentages can be tricky anyway, as you need to make clear to the reader what it represents. For example, the statement, “we found that putting up a poster in a bar reduced drink driving by 5%” can be interpreted as 5% of all drivers would not drink and drive or 5% of those who would have driven home after drinking will no longer do so.

Presenting Qualitative Results

By now you may be planning a PhD without numbers, but presenting qualitative results can face many similar issues. First and foremost, PhD studies using qualitative methods can generate huge amounts of data. Selecting and justifying what to include is a key task. Too much and you will drown your reader in data and lose your sense of narrative. Too little and you could be accused of cherrying picking the quoted results to suit you case. This is why you need an even stronger rationale for your presentation of selected or summarised results with qualitative studies than you do with quantitative studies, and this should have been presented to your reader as an integral part of the study design.

A second key issue for presenting qualitative results where interviews or other methods are used to gather data from individuals is confidentiality. If you have promised confidentiality, then you must be careful to preserve it, and this may involve more than just removing or changing names, taking care to ensure that there are no clues in quoted results that could identify the data giver. In some case, these considerations may preclude the inclusion of a complete set of results as an appendix or CD-ROM from the thesis.