We randomly choose a group of 1000 people, 500 smokers and 500 non-smokers from all over the world to investigate the effect of smoking on heart health. Is this a well-designed experiment?
We have two similar flowerpots which both contain a rosebud branch and the same type of soil. We put the first bud by the kitchen window and the second in a dark room. The rosebud in the first pot receives sunlight but the second rosebud does not receive any sunlight. We water both flowers every afternoon with the same amount of water.
After a few days, the first rosebud opens but the second rosebud does not open. Based on this experiment, we conclude that sunlight is necessary for rosebuds to open. Is this a well constructed experiment?
We choose a group of 100 runners. We are examining a new pile that claims to improve the running ability of runners. Assume that we give a pile to 50 runners and then ask all 100 runners to run. We compare the records of each runner before and after getting the piles. Suppose that these are the results of the experiments:
- The runners who do not have any piles have similar records comparing to their previous records.
- The runners who have the piles have better records comparing to their previous records.
Is this a well-designed experiment showing that the piles are really useful?
A doctor investigates the effect of salt on blood pressure. She randomly chooses 100 patients with high blood pressure and asks 50 of them to remove salt from their diet for a couple of weeks. Some of the patients have lunch in their office and are not able to bring their own lunches. To solve this problem, they decide to take leave for two weeks so that they can remove salt from their food during the period of the experiment.
Are the results of this experiment valid?
We randomly choose a group of 100 people at the main bus station of a town and ask them about fuel price.
- 75% of people believed that the government should raise the fuel price and invest it to develop public transportation.
- 15% of people believed that the government should not raise the price.
- 10% did not answer.
Is this a well-designed experiment?
We want know the opinions of students at a sporting event. At lunch time when all student are present, we randomly choose one person from each table and ask them poll questions. Are the results of this experiment valid?
We choose 100 people who have caught a cold. We are given a new pill that claims to be useful for cold. We randomly devote a number to each patient and ask them to come to the office one by one. We give odd numbered patients a real pill and even numbered patients a sugar pill. The patients are not aware if they are getting real pill or a sugar pill.
Is this experiment well-designed?