Psychometric studies are fundamentally important to ensure fair assessment processes in different contexts. Their use is not limited to a specific area or scientific discipline, being widely applied in fields such as Education. Regardless of the application context in which psychological characteristics are to be measured, it is necessary to employ techniques and procedures based on different psychometric traditions. Such techniques and procedures aim to generate evidence to support or refute the validity of the inferences made from the interpretations of the results produced by these instruments. Given the range of available techniques and procedures, it is crucial to contextualize the application of these different traditions, considering their substantive, methodological, and historical attributes and characteristics. In this sense, the present scientific investigation sought to focus on two distinct emphases: (1) the application of different psychometric techniques and (2) the exploration of methodological challenges, with the goal of providing evidence on both existing and newly developed psychometric instruments. To this end, the present thesis was organized into four articles. The first article aimed to gather evidence of measurement invariance for two mental health measures (depression and anxiety) in samples from three countries: Brazil, Spain, and Portugal. Overall, the results suggest the stability of the depression measure, while the invariance of the anxiety measure was found to be less stable. The second article focuses on gathering validity evidence for a recently developed instrument for assessing job penibility in the Brazilian hospital context. The findings indicate a more robust factorial structure with a single dimension, although contingent factors related to the theoretical proposition of a three-factor structure are discussed. In addition to the confirmatory investigation of the instrument s internal structure, an Item Response Theory (IRT) approach was implemented, revealing that the items, in general, presented moderate to high levels of difficulty. The third application refers to the use of Network Analysis techniques in the context of child development in the United States. The results can be interpreted within the framework of a stable network structure comprising six communities, which aligns with the theoretical proposal of the measurement instrument. Finally, the fourth study described in this thesis aimed to present opportunities and methodological challenges in handling large amounts of missing data within IRT models, also in the context of child development. Overall, the evidence presented in this thesis highlights the importance of adopting a substantively grounded approach to the application of different psychometric techniques and procedures. This is relevant not only for the investigation of internal structure but also for the development of new instruments, particularly when seeking to understand the intricate statistical relationships that emerge across different models. Furthermore, this study sheds light on current and future methodological challenges, especially considering the growing trend toward instruments that prioritize not only the needs of developers but also the experience of respondents.