Science Buddies: "Ask an Expert"

Hey,

So I just finished my science fair experiment, testing for heart rate recovery. I took the resting heart rate and systolic/diastolic blood pressure while they were resting, had them do exercise, and took their heart rate and blood pressure again 4 times in 1 minute intervals. 20 participants were randomly selected and tested and lots of their numbers were up and down. Realistically, they should have all dropped during the 4 minutes of resting, however, that was not the case for all-some data especially heart beats went up and down.

I was wondering what kinds of data analysis should I perform? T-tests? Since I can't not spot any direct correlation, I am thinking to divide data into 2 groups with normal BMI and higher BMI to prove the hypothesis that BMI affects the heart rate recovery and both blood pressure. I am still lost even though I read a few books about the statistics. How can I utilize my data to make it work!

Thank you for your time!
~Cybersix

A t-test would be a good start comparing the "normal BMI" group to the "high BMI" group, but with biological samples like yours they may have outliers - samples that don't fit the trend or support your hypothesis. Lots of different things can affect heart rate and blood pressure, including diet and even genetic factors. You could have a really skinny person with pulmonary hypertenstion because of genes. Have you just graphed the data out to see if you can find a trend by eye? If your samples aren't fitting your BMI hypothesis, maybe you can survey your participants and ask questions about their diet - how much fruit/vegetables they consume, salt, red meat, fish, fat, sweets and sodas (even diet soda has been shown to increase blood pressure!), etc.

Good luck!

If you aren't sure what type of statistical test to use, check this information out:
http://www.graphpad.com/www/book/choose.htm

Thanks for your reply! So now, I have another problem. I used the students t-Test on my data. I've split the 20 participants into 2 groups with BMI of <25 and >=25, comparing the data from both groups in t-test two sample assuming unequal variances at different time interval. If there were 10 participants, then the degrees of freedom would be 9, so the degrees of freedom for 20 participants should be 18 right (n-2)? The t-Tests I made in Microsoft Excel 2007 shown that the degrees of freedom ranges from 10-14, even though I put in 20 observations data. Can you please explain?