Final dissertation, worth 60 M-level credits, to demonstrate students’ capacity to conduct and report a major piece of research work

Figure 9 shows that most respondents (77%, n=36) agreed or strongly agreed that within their clinical area the best way to disclose trial results to participants would be to send the results or lay summary by post. Four respondents (9%) disagreed or strongly disagreed with this. The least popular of the suggested methods was sharing results by telephone call from a clinician; 29 (62%) PIs disagreed or strongly disagreed that would be the best option, although five (11%) agreed that this was the best method. A telephone call from a nurse or research coordinator was viewed as slightly more acceptable, but a face-to-face meeting with either a clinician or nurse was preferred to telephone contact. There was limited support for providing results via a web or journal reference to participants.

Figure 9: PI opinion about how trial results are best provided to participants

Respondents were asked to select up to two preferred methods for providing feedback to participants. Figure 10 confirms the provision of trial results or summary by post as the most favoured method (31 of a total 79 responses, 39%). The least favoured method was by face-to-face meeting with the clinician. No-one preferred to share findings via a telephone call from a clinician.

Figure 10: Personal preference(s) of PIs for return of trial results to participants

Figure 11 illustrates the responses to three attitudinal questions about returning trial results. More than half of respondents (57%, n=27) agreed or strongly agreed that 'disclosure of trial results should be at the request of the participant rather than being an automatic process' although 11 (23%) disagreed with this and nine had no particular view.

Figure 11: Attitudes of PIs towards the process of returning trial results

The majority of respondents (74%, n=35) agreed that 'participants should be told at the start of a trial how and when they can obtain the trial results'. Three PIs disagreed with this statement and nine neither agreed nor disagreed. Twenty eight (60%) PIs agreed that 'informing patients of trial results encourages their participation in future research', though 16 people (34%) had no particular opinion about this.

Some variation in opinion was noted between investigators of different specialties e.g. three (38%) oncologists/haematologists agreed/strongly agreed that disclosure of trial results should be at the request of the participant (i.e. optional) compared with six (75%) anaesthetists. Four (50%) oncologists/haematologists disagreed with this statement compared with one (13%) anaesthetist (N=8 in both specialties). Numbers were too small to assess statistical significance.

Function: Describe

Figure 10 is a pie chart, appropriate for expressing results as proportions of the whole population of respondents. Compare this with figure 9.

View linked text

Figure 11 is a further variation on the bar chart used. This time the number of responses to each option are presented as separate bars rather than as proportions of the total bar. This is slightly easier to read, but is only possible here because fewer options are represented than in Figure 9. Notice that the writer has chosen appropriate graphical representations for each data set. A program such as MS Excel can help with this.

View linked text

This extract is a good example of the use of figures and descriptive text together to describe complex results. The figures provide an immediate visual impression of the data, while the text adds explanatory detail. Figures should be numbered and titled, as they are here, and placed as close as possible to the text which refers to them.

View linked text

When describing statistical results it is good practice to qualify any statement of number ("most respondents") with the exact percentage and number, as the writer has done here in brackets.

View linked text

Figure 9 is a type of bar chart which allows several aspects of the findings to be represented visually in the same space. The different options for providing results are presented along the x axis in descending order of popularity. The different responses to each option are shown as proportions of the total response set by the different shaded regions within each bar.

View linked text

The writer describes the use of a differently phrased question to elicit respondents" opinions about providing results to participants after a trial. This allows the writer to confirm and strengthen an earlier finding. As with much good writing at Masters level, the description contains an element of analysis.

View linked text

Function: Analyse

The writer notes a pattern of difference which seems be related to another variable, in this case the specialism of the respondent. However, the writer is careful to point out that the numbers in this research study are too small to show any statistically significant correlation. It is still important to notice and comment on such patterns as they can provide an important indication to be followed up in future studies.

View linked text

General

This extract makes excellent use of graphs derived from statistical data to illustrate patterns in the data, supported by clear textual commentary.

View linked text

Quality: Voice

Description of statistical findings can be dry. The writer has made this section as interesting as possible by finding different ways to express the ideas "more popular" and "less popular".

View linked text