Question

1. The authors claim that random assignment is “the key to eliminating all possible confounds in a true experiment.” however this is not correct. Can you think of a type of confound that would not be eliminated by random assignment? Can I please get an example?
2. Is it possible to use a ratio scale to measure a psychological construct like Anger or Happiness? If so, give an example. If not, explain why not.

Thank you for your help!

Answer 1

1. Random assignment can eliminate confounds that arise due to differences between individuals. However, it cannot compensate for a poor experimental design. An experimental design may have significant flaws. Even when the participants are randomly assigned, the flaws might not be eliminated. For example, in an experiment designed to estimate the effects of music on learning, the two conditions (one experimental and one control) might be administered at different times of the day. Suppose the group that receives music is administered the condition during morning and the control group during evening, the differences that arise between the two groups can be caused by the time of the day and not solely by music.

2. A ratio scale has a true zero point. It has equal intervals between two measurements. Psychological constructs like anger and happiness cannot be measured using this scale. A true zero point means the absence of an attribute. No one can possibly have 0 of a psychological construct. One cannot have zero happiness, aggression or intelligence. Moreover, if an individual has a score of 2 and another one has a score of 4, we cannot objectively conclude that the latter's happiness is twice that of the former's. Hence, a ratio scale cannot be used for such constructs.