# 🐅2.78🐣 Empirical Methodology [Sections 2.7-2.8] ## 2.7 Substantial Meaning to Latent Variables The key to our empirical approach is assigning substantial meaning to latent variables in the hierarchical model. Unlike purely statistical constructs, our parameters have direct interpretations: - **φ (promise level)**: The entrepreneur's committed success probability - **μ (aspiration)**: The mean level of ambition across ventures - **τ (concentration)**: The precision of the promise, inversely related to flexibility - **c (complexity)**: Environmental difficulty, measurable through operational metrics - **i (integration cost)**: The cognitive and organizational cost of processing new information This substantive grounding enables identification. We can measure: - τ through the variance in pivots and strategic changes - c through the number of stakeholders and technical dependencies - i through organizational size and decision-making speed The hierarchical Bayesian framework naturally accommodates this structure, allowing us to estimate group-level parameters while accounting for individual variation. ## 2.8 Essential Heterogeneity and Hierarchical Structure Essential heterogeneity—systematic differences between founders that affect outcomes—evolves through our hierarchical structure from group to individual levels: **Group Level (Industry/Ecosystem)**: - Mean aspiration μ̄ varies by industry (biotech > software) - Baseline complexity c₀ differs across sectors - Information costs i reflect ecosystem maturity **Individual Level (Founder/Venture)**: - Founder-specific τ reflects experience and confidence - Venture-specific adjustments to c based on technical choices - Learning rates vary with individual i This group-individual structure explains patterns invisible at either level alone: - Why certain industries favor high τ (pharma) vs. low τ (software) - How ecosystem effects shape individual choices - When group norms override individual optimization The evolution from essential heterogeneity (founder differences) to hierarchical variation (group-individual structure) provides the empirical foundation for testing our theoretical predictions.