Understanding the role of different age groups in disease transmission is crucial for designing effective intervention strategies. Age-structured epidemic models capture these dynamics through contact matrices, which describe interactions between subpopulations. However, empirical contact estimation is inherently prone to measurement noise, as survey responses vary systematically across age groups and are collected through heterogeneous channels. This imprecision introduces substantial variability into epidemic predictions. In this study, we present the Age Group Sensitivity Analysis (AGSA) framework for quantifying how structural perturbations in age-specific contact patterns propagate to epidemic outcomes. AGSA integrates age-stratified epidemic models with Latin Hypercube Sampling (LHS) and Partial Rank Correlation Coefficient (PRCC) analysis to perform a systematic sensitivity assessment of age-specific interactions. A key novelty of our approach is a sensitivity aggregation technique that attributes the overall dispersion in epidemic outcomes to individual age groups. By identifying the groups whose contact variations contribute most to model variability, AGSA highlights where refined empirical data are most urgently needed. This provides a principled basis for targeted data collection efforts, thereby constraining epidemic forecasts and supporting the robust evaluation of age-specific public health interventions.