Name order effects in measuring adolescent social networks using rosters

Area of Science:

• Social network analysis
• Sociometry
• Psychological research methods

Background:

• Order effects, where list position influences responses, are a concern in social network data collection.
• Previous research suggests randomizing or sampling names to mitigate potential biases in nomination data.

Purpose of the Study:

• To model and assess order effects as biases in nomination choices within social network data.
• To investigate the mechanisms underlying these potential biases.
• To identify factors that may amplify or reduce order effects in survey situations.

Main Methods:

• Utilized observational and experimental methods to assess nomination biases.
• Analyzed data from four middle school cohorts across six data collection waves.
• Examined specific aspects of the survey environment as potential moderators of order effects.

Main Results:

• Found limited evidence of significant order effects across eight sociometric questions.
• Data from multiple cohorts and time points did not consistently support the presence of strong order effects.
• Identified survey situation factors that can influence the magnitude of order effects.

Conclusions:

• Order effects may not be as prevalent in sociometric data collection as previously assumed.
• Understanding moderating factors is crucial for accurate detection and mitigation of biases.
• Findings provide practical guidance for optimizing survey design in social network research.