A student is conducting their science experiment on the effect of caffeine on dogs. He has 3 groups of test subjects. The 1st group of dogs receives plain water. The 2nd group of dogs receives 10 mg of caffeine each, and the 3rd group receives 50 mg of caffeine each. He will measure their activity levels by recording how long each dog runs without stopping, after giving them the pills. What could be evaluated in this experiment to ensure reliability?