Does alcohol intake increase the reaction time of a driver? In this report, I will analyze reaction times of 20 drivers as recorded before and after they had two beers. The table below contains the reaction times and their differences.
| SubjectID | Before | After | AfterMinusBefore |
|---|---|---|---|
| 2 | 2.96 | 4.78 | 1.82 |
| 13 | 3.16 | 4.55 | 1.39 |
| 4 | 3.94 | 4.01 | 0.07 |
| 16 | 4.05 | 5.59 | 1.54 |
| 17 | 4.42 | 3.96 | -0.46 |
| 20 | 4.69 | 3.72 | -0.97 |
| 6 | 4.81 | 5.34 | 0.53 |
| 5 | 4.85 | 5.91 | 1.06 |
| 10 | 4.88 | 5.75 | 0.87 |
| 3 | 4.95 | 5.57 | 0.62 |
| 18 | 4.99 | 5.93 | 0.94 |
| 19 | 5.01 | 6.03 | 1.02 |
| 9 | 5.15 | 4.19 | -0.96 |
| 12 | 5.26 | 7.23 | 1.97 |
| 8 | 5.33 | 5.84 | 0.51 |
| 15 | 5.49 | 5.25 | -0.24 |
| 11 | 5.75 | 6.25 | 0.50 |
| 1 | 6.25 | 6.85 | 0.60 |
| 7 | 6.60 | 6.09 | -0.51 |
| 14 | 6.65 | 6.42 | -0.23 |
There are only 20 data points. Therefore, we have to verify that the data is normally distributed before proceeding with the test. We can use a histogram to analyze the distribution.
From the histogram above, we can see that the data is normally distributed. Additionally, from the qqplot below, every data point falls in the 95% confidence band indicating that the data is normally distributed.
We’ll start by stating the null and alternative hypotheses for our test. \[H_0: \mu=0\] \[H_a: \mu>0\] I will now run a paired sample t-test for the data.
##
## One Sample t-test
##
## data: joined_table$AfterMinusBefore
## t = 2.6031, df = 19, p-value = 0.008734
## alternative hypothesis: true mean is greater than 0
## 95 percent confidence interval:
## 0.1690516 Inf
## sample estimates:
## mean of x
## 0.5035
As we can see, the t-score is 2.6031 which gives a p-value of 0.008734.
\[p < \alpha \ \ \ \ \therefore\ \ we \ \ reject \ \ the \ \ null\]
At a 0.05 level of significance, there is enough evidence to conclude that there is an increase in the average reaction times of drivers after alcohol intake.