An AI can estimate job satisfaction effects as well as surveys, given enough details.

Standard surveys often leave out important details about people's jobs and daily lives. To fill these gaps, researchers created AICOME, or AI Contextual Measurement. This framework tests whether artificial intelligence can build missing survey measures that capture effects for both individuals and broad groups. Instead of simply predicting individual survey responses, AICOME creates an AI measure for each person, calculates an overall group average, and finds how much each person differs from that average. This setup allows researchers to study patterns between groups as well as within groups. Testing the framework on occupational survey data, the researchers found that AI can recover much of the contextual information found in real survey questions if rich job and background details exist. For example, AI measures of weekly hours successfully matched real survey findings showing negative links with satisfaction. However, AICOME performs worse when data is scarce or when several related ideas are missing at once.

Why it matters

Researchers can use AI to fill gaps in surveys, making studies more efficient without new data collection. This helps social scientists explore topics where survey data is missing, affecting how we understand jobs and well-being.

Questions answered