Over the last few decades, rapid technological advancements have raised challenging and pressing questions for education. Among them, the critical competencies required for a future-ready generation are continuously debated. Computational thinking (CT), entailing a set of cognitive skills important for programming and problem-solving, has emerged in the literature as such an essential competence domain. As a newly emerging construct, there are two fundamental questions in the CT literature needed to be illuminated: what is the nature of CT and how is it evidenced? In other words, is CT a domain-specific skill, a domain-general skill, or a competency reflecting a combination of both? These questions are explored through three empirical studies, which address different aspects in conceptualization and assessment undertaken by different methodological approaches. Throughout the thesis, I redefine the understanding of CT from a traditional, common domain-specific perspective (e.g., programming-centric view) by proposing a hybrid account that perceives the construct as a complex competence domain. In doing so, my theoretical proposition of CT competency bridges domain-specific and domain-general perspectives, arguing it encapsulates both programming and problem-solving skills. Starting with conceptualization, Study 1 investigates teachers’ views using a text mining approach, shedding light on the nature of CT and appropriate assessment design. The results supported my design of a technology enhanced assessment of CT competency, Computational Thinking Challenge (CTC). The CTC adopts an integrative assessment design, which includes multiple contexts, formats, and problem types in reflecting the diversity of CT competency. Hence, it is critical to address the issue of whether the diversity of the tool impacts its quality. For this, Study 2 conducts a rigorous psychometric evaluation using a Rasch model from an item response theory approach. Extending from this, Study 3 examines the structure of CT competency, using a multidimensional item response theory approach, to investigate the structural validity and item quality of the CTC. In addition, a bi-factor model is specified to test my proposed theoretical model that consists of a general CT competency factor and two specific factors in programming and problem-solving. Together, the studies provide supporting evidence to my theoretical proposition and help elucidate the nature of CT as a competence domain. The studies also demonstrate the promise of adopting an integrative assessment design in evidencing CT competency through the CTC. In short, this thesis addresses questions that bridge conceptualization and assessment: (1) What is CT competency? (2) How should it be evidenced? and (3) Is the measurement method reliable and valid?