Pace as the Fatigue Measurement

As I promised, I measured my pace and correlated it with my subjective fatigue rating over 4 week period. The data looks as follows:

date	duration	fatigue
9/17 9:02	19.71	5
9/17 9:45	20.25	4.8
9/17 11:30	19.01	5.1
9/17 11:30	18:53	5.1
9/17 12:02	18.69	4.9
9/17 12:02	19.37	4.9
9/17 18:08	17.56	5.2
9/18 10:46	18.69	5.1
9/18 12:07	18.9	5
9/18 14:45	19.61	5
9/18 10:48	16.75	5.3
8/18 14:02	18.09	5
8/18 20:13	17.61	5.2
8/19 11:40	17.94	5.2
8/19 15:35	18.78	5
8/21 11:51	17.61	5.1
8/21 12:44	18.64	5
8/21 14:12	18.35	5
8/22 9:50	17.9	5.2
8/23 9:33	17.39	5.2
8/23 9:58	17.2	5.3
8/23 13:04	18.55	5
8/23 11:58	17.81	5.1
8/25 8:46	18.56	5
9/25 10:58	17.45	5.2
9/25 23:02	17.2	5.3
8/27 16:44	18.12	4.9
9/28 16:57	17.42	5.3
9/29 10:38	17.96	5.1
9/30 12:59	19.28	4.8
10/2 10:34	16.78	5.3
10/3 13:51	17.83	5
10/4 10:48	17.21	5.2
10/4 13:09	18.41	5
10/4 18:02	18.28	5
10/4 20:08	16.45	5.3
10/5 9:37	16.49	5.3
10/6 11:24	16.82	5.2
10/7 11:25	17.77	5
10/7 20:43	15.87	5.3
10/8 11:43	16.69	5.3
10/8 20:20	18.08	5
10/8 11:23	16.96	5.2
10/8 12:20	17.45	5.1

The duration is the amount time in seconds that took for walking 29 steps. (I meant to do 30 steps, but I started counting from 1 instead of 0, so I ended up with 29 steps). The fatigue is my subjective fatigue rating as usual. The resulting scatter plot between the duration and fatigue looks like this:

The walking speed is nicely correlated to the fatigue as you can see. (The Y scale looks funny, that's what R does when you do as.numeric(pace$duration), but it is properly scaled). The correlation comes out to be 84%.

The problem with this study, of course, is that I am both the investigator and the subject: there is no blinding whatsoever. As result, it is highly susceptible to bias. But the result is at least encouraging. Next, I will write an android app to automatically measure my pace without my knowing it so that I'll be blinded to it, and see if that correlates to the fatigue level as well.

If you want to try this yourself, do the following to minimize the bias:

1. Declare your fatigue first.
2. Walk fixed number of paces and measure the time.
3. Focus on how your leg feels. Make sure that you are walking as fast as possible without any straining at all.
4. Whatever you do, do NOT revise your fatigue rating after the walk. It is tempting, as you realize while walking that your fatigue is not what you thought it was. But the error is random and therefore will cancel out on average.