# Building Models in Two Worlds: From Latent Constructs to Behavioral Signals

Differences in modeling between survey data and behavioral data

- **Source:** [Towards Data Science](https://towardsdatascience.com)
- **Published:** 2026-07-13
- **Topic:** [Data & AI](https://brifing.si/c/data-ai)
- **Original article:** <https://towardsdatascience.com/building-models-in-two-worlds-from-latent-constructs-to-behavioral-signals/>
- **On Brifing:** https://brifing.si/s/047cffdd-4659-4e2f-8098-6c5325b5f128

## Introduction

Modeling requires different approaches when working with both survey data and behavioral data. This article explores the fundamental differences between these two worlds.

## Survey Data

Survey data is based on participants' statements and often contains latent constructs. These constructs must be reconstructed from indirect evidence.

## Behavioral Data

Behavioral data records users' actual actions and can be directly analyzed. This data is readily available in databases.

## Technology Acceptance Model

The Technology Acceptance Model explains users' adoption of technology. This model focuses on factors such as perceived usefulness and ease of use.

## Structural Equation Modeling

Structural Equation Modeling uses both measurement and structural models. This method allows for modeling measurement errors.

## Latent Constructs

Latent constructs are created from survey question data. These constructs represent shared variances and are cleaned from measurement errors.

## Measurement Errors

Measurement errors are the noise in survey responses. These errors can affect regression coefficients.

## Statistical Power

Each latent construct requires statistical power. Therefore, sample size and the number of indicators are important.

## Theory and Data

Theory is more important than data in modeling. Theory determines which variables are included in the model.

## Results

Modeling results can emerge in unexpected ways. Some variables may be less effective than expected.

## Future

Modeling techniques are shaping the future of data science. Building effective models in both worlds brings success.

## Importance

Understanding the differences between survey and behavioral data is critical for building more accurate models.

## Sources covering this story

- https://towardsdatascience.com/building-models-in-two-worlds-from-latent-constructs-to-behavioral-signals/ — 

---

_Summary written by Brifing from the original reporting. Copyright in the original article belongs to its publisher; read the full piece at the link above._