Emotion pervades all spheres of human life. However, it is surprising to notice the scarcity of research tapping into the role of emotion in career development, vocational behavior (Hartung, 2011; Kidd, 1998) and learning, especially given its interpersonal and relational nature. One of the most limiting factor in the study of emotion has been its elusive nature, making it difficult to define and measure it (Kleinginna & Kleinginna, 1981; Oatley & Johnson-Laird, 1987). Following the multitude of conceptual definitions, a multitude of different approaches in emotion measurement emerged in the literature encompassing physiological, behavioral and self-report techniques. Accordingly, measuring emotion can become a very tedious task for researchers (Mauss & Robinson, 2009; Meiselman, 2016). In this presentation, we propose to discuss several issues regarding the use of self-report measures of emotion. Self-report measures have a series of advantages. For example, they are easy-to-use and participants may more easily accept to fulfill a questionnaire compared to invasive measurements such as electrodes. They also respects structural and psychometric properties such as valence and activation. Finally, they allow for multi-dimensional measurement such as frequency or intensity. However, research evidence has also cautioned about the use of self-report measures. For example, self-report measures only capture conscious emotions. In addition, participants can also distort (consciously or unconsciously) their responses in order to present themselves in a more positive way (Welte & Russell, 1993). Some participants may as well be unable to report their emotional experience (e.g. alexythimia, Lane, Ahern, Schwartz, & Kaszniak, 1997). Finally, it also exists some time-related bias when we ask participants to report their emotional state. Robinson & Clore (2002) have for instance demonstrated that current emotional experience tend to be more valid than self-report concerning future or past. Unfortunately, self-report measures cannot capture all aspects of emotional experience. Despite its limitations, self-reported measures of emotions are widely used in research. In this presentation, we present an illustrative study that address some limitations of self-report measures with a mixed-method design and a person-centered approach by investigating experienced emotion of adult learners at the beginning of their adult education program. Extending on Ekkekakis & Zenko (2016) recommendations suggesting to (1), carefully consider the definition of emotion (2), choose the most appropriate theoretical framework for the chosen construct and (3), select the psychometrically strongest measure based on the chosen theoretical framework, we propose additional recommendations. Specifically, we propose that researchers should pay attention to the following questions : (1) how do they conceptualize emotions (state/trait, anticipative/felt, general/specific affect(s)), (2) how do they model emotion in their study (goal- or event-related, dependent/independent variable, within/between variability) and (3) feasibility (diary study, survey length, experience sampling method, population of interest, cross-sectional/longitudinal designs).