Research Design in Psychology Possible sources of headache
What may go wrong in research Conceptualization problems Design problems Measurement design (= operationalization) problems Measurement implementation problems Statistical problems Interpretation problems
Concepts of Design: NOT to be confused with Measurement (PSYC410) Reliability Validity – internal Validity – external
Potential design problems The worst and the most problematic: IV does not vary Other events/experiences may interfere Between WHAT and WHAT Selection of the sample and groups – representativeness / attrition / non- participation Problems with extreme groups Problems with measurement The way you measure or the fact that you measure it
Example: No variability Divorced parents Stability of relationships
Example: solution Whether there is a parental divorce Stability of relationships
Example: Instrumentation Reflection on parental divorce and feelings of insecurity Whether there is a parental divorce Report of stability of relationships Actual stability of relationships
Example: History IV HISTORY DV
Example: History (other “IV’s”) Whether there is a parental divorce Stability of relationships Experience with relationships throughout adolescence
Example: history (Maturation) Whether there is a parental divorce Stability of relationships Ability to interpret parental discord and evaluate it in perspective
Example: History (Social and personal events) Whether there is a parental divorce Stability of relationships Economic difficulties, residential moves, changes of school, changes of social relationships
Example: History that occurs prior to the IV (third variable) Psychological well being of parents prior to divorce Whether there is a parental divorce Stability of relationships
Example: Sample (Biased selection of subjects) Whether there is a parental divorce Stability of relationships [KU]
Example: (explanation)Biased selection of subjects Family functioning OSS Scores Whether there is a parental divorce Stability of relationships [KU]
Example: Biased selection of subjects (regression towards the mean) Whether there is a parental divorce Stability of relationships [Clinical Sample]