Research Adequacy (2.6)

2.6.1. Research is considered ethical only if it meets appropriate standards of adequacy.

2.6.2. AUTEC recognises that different research paradigms may inform the conception and design of projects. It adopts the following minimal criteria of adequacy:

• The project must have clear research goals;
• Its design must make it possible to meet those goals;
• The project should not be trivial but should potentially contribute to the advancement of knowledge to an extent that warrants any cost or risk to participants.

Note that in the case of student projects, it is appropriate to take into account the contribution of the work to the student’s own education.

2.6.3. Quantitative research

2.6.3.1. In deciding the number of participants required for a quantitative study there must be a balance between:

• the effective use of participants so that their time and effort is not wasted;
• the measurements that are being made all being necessary and of adequate accuracy and precision;
• the minimum number of participants taking into account dropouts; and
• the likelihood of verifying that a difference or change actually exists (power of study) and that the change expected will have biological significance.

2.6.3.2. In quantitative research, the hypothesis must be clearly stated and be specific about what variables are being measured and the units of measurement.

2.6.3.3. How the power calculation was done must be described fully so that it can be verified. i.e.

• the clinical effect or difference hypothesized to be seen;
• variability on which the power calculations are based - i.e. standard deviation of the measurement taking into account all variation;
• the statistical test that will be used;
• the sources of the information.

e.g.  it is known (see reference below) that without treatment the mean and standard deviation of a waist measurement in a previously studied group is 108±2cm.  The lifestyle intervention will be considered successful if after treatment the waist measurement difference between a control and treatment group is 2cm.  The question is “How many should be in each group to be able to distinguish this difference with 90% power and a probability (P) of 0.05?”

The test to be applied is an unpaired t test.
The answer is that there should be 23 in each group i.e. 46 total.
Racette SB, Weiss EP, Obert KA, Kohrt WM, Holloszy JO. Modest lifestyle intervention and glucose Obes Res. 2001 Jun;9(6):348-55.
i.e. measurement description
actual value and units of measurement and standard deviation
the change expected
power value
P value
Test to be used
And the final number
Reference or pilot study where information came from Action research: For guidance on action research projects, please refer to Appendix S.