PSYC 3380 Lecture Notes - Lecture 3: Internal Validity

• School: University of Guelph
• Department: Psychology
• Course: PSYC 3380
• Professor: Jeffrey Spence

Document Summary

Designs without control groups: one time observation. *most common design because it is easiest: one group pretest-posttest design. O1 o2: could be strengthened by adding more pretests and posttests. O1 o2 o3 o4 *tracking someone over 4 periods in time. Note: the more x"s and o"s added to the design, the stronger/more complex the design is. O1 o2 o3 o4 o5 o6. Nr o: nr= non-random assignment, x= treatment, o=observation/measurement, naturally occurring groups, weakness= no pretest. Makes it susceptible to all forms of selection threats. We don"t know if these groups are actually different. If we had a pretest, we could start to rule out if the groups were different in the first place. Nr o1 o2: improvement on posttest only design, can test for group equivalence before treatment by taking a measure of our variables before the test, equivalence by not be stable: Selection-regression threats: if we selected based on extreme scores.