Using Sensors in Organizational Research-Clarifying Rationales and Validation Challenges for Mixed Methods

Area of Science:

• Social Sciences
• Behavioral Sciences
• Organizational Sciences
• Computer Science

Background:

• Sensor-based big data are increasingly used but lack systematic validation.
• Sensor data are complex, constructed sources, not neutral views of reality.
• Existing research is fragmented across disciplines, hindering a unified approach.

Purpose of the Study:

• To address the lack of systematic validation for sensor-based data collection.
• To debunk the myth of raw data and emphasize data construction.
• To provide recommendations for designing mixed-methods studies using sensors.

Main Methods:

• Literature review across social, behavioral, and computer science.
• Conceptual analysis of sensor data validation.
• Discussion of interdependence between sensors, constructs, and context.

Main Results:

• Sensor data validation requires considering sensor types, research constructs, and context.
• Incompatibility between granular sensor data and traditional static data is a key limitation.
• Failure to apply social science measurement criteria can lead to insignificant results.

Conclusions:

• A systematic approach to sensor validation is crucial for reliable social science research.
• Integrating sensor data with traditional methods requires careful consideration of temporal and granularity mismatches.
• Recommendations are provided for designing robust mixed-methods studies incorporating sensor data.