Child Development Lecture: Methods Replication

Scientific Method

• the process of testing ideas about the world by

  • setting up situations that test our ideas

  • making careful, organized observations

  • analyzing whether the data fit with our ideas

• observation, ask a question, formulate a hypothesis, experiment to test the hypothesis, record data/analyze, draw a conclusion

Confounding Variables

• Variables other than the independent variable/s that may also potentially explain the outcome

• Rule these out and control for them

Correlational Designs

• to determine how (if at all) variables are related to one another

  • correlation: the association between two variables

    • positive and negative relationships

Correlation ≠ Causation

• Direction-of-causation problem

  • it is not possible to tell from a correlation which variable is the cause and which is the effect

• Third-variable problem

  • a correlation between two variables may arise from both being influenced by some third variable

Experimental Designs

• Allow inferences about cause and effect

• Make systematic changes in certain factors and observe changes that occur in other factors

Quasi-Experimental Design

Quantitative research approach used to find cause-and-effect relationships when you cannot randomly assign participants to groups

Cross-Sectional Design

• Compare different children of different ages, at ONE point in time

  • pros:

    • fast, cheap, low risk of dropout

  • cons:

    • can’t track change in an individual

    • cohort effects: a group of people who share a common life event (like year they were born) or when they lived through a major historical event—develop shared traits, behaviors or knowledge

Longitudinal Design

• follow the same children over time

• Pros

  • real individual change is tracked

  • no cohort confound

• Cons

  • slow, expensive, lose participants

Microgenetic Design

• Study children repeatedly, over a short window, while a specific skill is actively changing

• Pros

  • captures the actual process of change, not just before-and-after snapshots

  • rich, detailed picture of how change unfolds

• Cons:

  • only works if you can identify children right on the cusp of specific change

  • intensive to run

  • Findings may not generalize beyond that narrow transition window

Between-Subjects Designs

• Different children, different conditions. Each child sees only one

  • Pros

    • no carryover between conditions

  • Cons

    • individual differences between group can muddy the comparison

    • needs more total participants

Within-Subjects Design

• The same children experience every condition

• Pros:

  • each child is their own control

  • Needs fewer total participants

• Cons

  • order effects

  • fatigue, practice, boredom

  • carryover between conditions

Replicability

• do independent studies, using the same method, get the same results?

• replication is the only way to tell the difference if the original results were true

Replicability Crisis

• methodological challenge in which researchers cannot repeat or duplicate the results of published scientific studies

Many Babies Project

• Collaborative project for replication and best practices in developmental psychology research

• goal to bring researchers together eto address difficult outstanding theoretical and methodological questions about the nature of early development and how it is studied