Effect sizes and effect size benchmarks in family violence research

Area of Science:

• Psychology
• Sociology
• Criminology

Background:

• Scholarly journals increasingly require effect size (ES) estimates, but authors lack clear interpretation guidance.
• Existing ES benchmarks, like Cohen's (1988), may not be suitable for all research areas.
• Inconsistent ES interpretation can affect the perceived meaningfulness of research findings.

Approach:

• The study discusses common ES estimates and Cohen's (1962) method for developing benchmarks.
• New ES benchmarks were established for family violence (FV) research using data from Child Abuse & Neglect (2021).
• ESs were analyzed across FV subspecialties to provide contextualized benchmarks for power analyses.

Key Points:

• The medium ES benchmark for general family violence research (d = 0.354) is smaller than Cohen's general recommendation (d = 0.500).
• Context-specific ES benchmarks are crucial for accurate interpretation of findings within specialized research domains.
• Developing discipline-specific ES benchmarks presents unique challenges.

Conclusions:

• This review aims to enhance understanding of ESs and their interpretation for professionals in family violence research.
• The findings provide valuable benchmarks for planning future FV research and conducting power analyses.
• Accurate ES interpretation is vital for understanding the meaningfulness and impact of research outcomes.