variation between the individuals (participant variables) in each group may affect the DV measurements making it look as if the IV has had an effect when it actually hasn't
What could you control when using independent groups?
After the participants have been recruited they should be randomly assigned to their groups. Ensuring groups are similar average