Sequential Bayes Factor designs in developmental research: Studies on early word learning.

Area of Science:

• Cognitive Development
• Developmental Psychology
• Bayesian Statistics

Background:

• Developmental research often faces challenges with small sample sizes and inconclusive results.
• Infant participant recruitment and response variability complicate traditional study designs.
• Sequential testing offers a potential solution to these methodological hurdles.

Purpose of the Study:

• To implement a Sequential Bayes Factor (SBF) design to address low sample sizes in developmental research.
• To examine infant sensitivity to mispronounced words, word-object association learning, and the mutual exclusivity assumption.
• To evaluate the efficiency and effectiveness of SBF designs in developmental studies.

Main Methods:

• Applied the Sequential Bayes Factor (SBF) framework to three developmental findings.
• Tested an initial sample of 20 infants per study, incrementally increasing sample size.
• Computed Bayes Factors with each additional participant to assess evidence for hypotheses.

Main Results:

• Achieved moderate evidence for infant sensitivity to mispronounced familiar words with a reduced sample size.
• Failed to replicate findings for cross-situational word learning; data favored the null hypothesis.
• Concluded infants did not recognize trained word-object associations in the cross-situational learning task.

Conclusions:

• Sequential Bayes Factor (SBF) designs can provide robust evidence with smaller sample sizes in developmental research.
• The SBF approach facilitated the rejection of a previously reported finding in infant word learning.
• Highlights the utility and challenges of implementing SBF designs for developmental researchers.