A pharmaceutical company is testing a new drug for diabetes patients. Due to various side-effects, they start with a very small sample, ten subjects for their first trial. The subjects take the medication with their evening meal. After the meal, they test their blood sugar (mg/dL). The hope is that the blood sugar levels will be lower with the new medication. Five subjects were kept in a control group and five took the medication, randomly assigned. Here are the data. 

1. Name and describe the design of this experiment. Are you satisfied with this design, or could you suggest an improved design? 

2. Make a graphical display to compare the two groups. 

3. Using your graph and summary statistics, write a short description of the two groups and draw a conclusion about the effectiveness of the medication by calculating a statistic. With such small sample sizes, we might wonder if this difference is actually caused by the difference that could have occurred simply due to the "luck of the draw". If you will study a method to compare two groups like these. However, for this investigation, we are going to use a simple randomization technique to examine if the difference we observed is truly due to the treatment, or if it could be due to chance. Make ten slips of paper (3x5 cards work well) with each of the ten blood sugar levels written on them. Shuffle. Randomly assign five cards to the treatment group and five to control. Calculate the mean difference between the groups. Repeat. 

4. Graph the mean differences you obtained. You should have 25 such differences. Indicate on the graph the difference from the actual experiment lies. 

5. What do you conclude? Is the difference observed in the experiment one that could have happened by random chance? Or does it appear that the difference is unusual and was due to the medication being administered to the five subjects?

treatment 185 163 125 151 101 control 190 175 178 135 201