Statistical Analysis

In your proposal you are asked to describe statistical tests for significance that you will perform on the data that you collect in the evaluation of your device. Below are two examples of such a description. Note that each example identifies the following.

1. What data will be collected.
2. How many times the measurement will be repeated.
3. The null hypothesis to be tested.
4. The alternative hypothesis.
5. The value of the probability statistic that is considered significant.

Example description of Student’s T Test

Pressure drop will be recorded at the given flow rate 5 times for the two stenosis lengths. The null hypothesis, that pressure drop does not depend on stenosis length will be tested by a one-tailed Student’s T test, under the alternative hypothesis that pressure drop is larger for longer stenosis. The result will be considered significant for p values less than 0.05.

Example description of Pierson’s Correlation Coefficient

Pressure drop will be recorded at 7 flow rates, ranging from 2 ml/sec to 15 ml/sec. A linear least squares fit of these data will be performed, and the Pierson’s correlation coefficient from this fit will be evaluated to test the hypothesis that pressure drop depends on flow rate in the range of flow rates examined. The null hypothesis is that pressure is not a function of flow rate. Results will be considered significant for p values less than 0.05.

Steven A. Jones

Senior Design (BIEN 400, 402 and 404)