Original Hypothesis:
I am a veteran with PTSD. Due to this, I do not have a regular sleep schedule. I have noticed that when I sleep very little, or not at all, I seem to make more mistakes, particularly early in the morning.

I will record the amount of sleep, in minutes, that I get each night. Each morning, within one hour of waking, I will take a typing accuracy test from the same source. I will record the number of errors in the typing. I expect that there will be a negative relationship between the hours of sleep I get and the errors that I make; by which I expect to have less errors with more sleep.


At first, I did this a bit begrudgingly. I am not a morning person, and forcing myself to do this on a nearly daily basis was frustrating at first. After a week or so, however, this became a part of my morning ritual. I used the following website to conduct my typing tests: http://www.typeonline.co.uk/typingspeed.php
I did this for a total of 30 measured days, and I have included the raw data below as a starting point:

Amount of sleep*
Mistakes made
*Measured in hours, rounded to the nearest 1/4 hour
2.25
11

6
2

4
5

5
4

6.5
3

5
4

3.75
7

4
6

2.25
10

8
0

7
1

3
6

2.5
9

2.5
10

3.75
8

5.75
5

8
1

6
3

4
8

2.25
8

7.5
2

3.75
6

6.5
3

3.25
8

4.5
5

8.25
1

5.5
6

6
2

4
8

3.25
8

4.8000
5.3333
Averages (rounded to 4 decimal places)
Sample size: 30



scatter plot.jpg
Correlation coefficient (r): -0.93443488642444
Coefficient of Determination: 87%
Linear Least Squares Regression Line Equation Calculation:
Slope(m) -0.684
Y – intercept 8.435
Least Square Regression Line Equation Y -0.684x + 8.435
Results:
Initially, there could have been a lurking variable in this test, as I intended to type the same paragraph daily. This could have resulted in memorization of the text, which could have improved my error rate. I removed this variable by using a site that generated a random test of the same approximate typing difficulty each day.
The correlation coefficient and the resulting coefficient of determination of 87% tell me that there is an 87% correlation between the amount of sleep that I get per night and the amount of errors that I made in the typing tests within an hour of waking up. This is a very significant correlation.
This confirms my initial hypothesis, in which I inferred that the amount of sleep that I get would affect my ability to type correctly in the morning.