Teacher self-efficacy has predominantly been explored using quantitative instruments such as Likert scales-based questionnaires. Several researchers have questioned these methods, suggesting they offer only a limited view of the concept. This paper considers their claim by exploring the self-efficacy of UK secondary science teachers participating in a two-year professional development programme using both traditional quantitative scales and qualitative methods, including interviews and lesson observations. The findings support the suggestion that traditional quantitative scales do not fully capture teacher self-efficacy and highlight inconsistencies between self-efficacy assessments through the different research approaches. We argue that to achieve a more complete and comprehensive picture of teacher self-efficacy, it is essential that traditional quantitative approaches are better triangulated and integrated with other sources of data, in particular lesson observations. We offer an emerging approach of how qualitative data sources might be used to develop this comprehensive picture.