Sala [1] expressed concerns that the findings from our study published last year in this journal [2] were biased due to limitations in our measures of affect and emotion. Responding to the points raised by Sala [1], we will first address concerns that the measures employed did not address all relevant emotions for individuals with eating disorders (EDs). We will then discuss considerations for future research, extending suggestions made by Sala [1].

First, we agree with Sala [1] that our measures (PANAS, POMS, STAI) may not capture all relevant emotion states for individuals with EDs. It is true that various emotions were not assessed by the PANAS, POMS, and STAI. However, assessing all theoretically-relevant emotions was not the original intent of this investigation. Instead, these measures were used for two primary reasons. First, instruments were selected for high reliability and validity in evaluating a range of affective states and emotions in the study sample, in addition to their apparent utility in other ED studies. Second, these measures assess emotional domains that we believed were most relevant given the prior literature. Nonetheless, we agree that research evaluating associations between additional emotions and EDs is needed; we hope this paper will help to spur such work. In particular, future investigations should evaluate distinctions between specific negative and positive emotions, and evaluate differential associations between ED risk and basic (i.e., fear, disgust) or complex (i.e., guilt, shame) emotions.

We also concur that research should assess emotional change during mealtimes. Assessing change in the context of eating may provide valuable insight into emotional changes that treatment-seeking individuals experience during mealtimes in clinical settings. Although our study was not designed to assess change throughout mealtimes, we agree that such an investigation would contribute to the literature. Future work should also evaluate change in affect during other types of eating episodes (e.g., snacks, binge eating episodes) and in different surroundings (e.g., naturalistic environments).

Overall, we agree that future research should investigate a wide range of emotions. However, the assertion that our results are biased due to lack of assessment of all relevant emotions is not necessarily warranted, considering the risk for Type I error, had we examined additional emotions given our sample size, and the fact that the measures used in our study were specifically selected a priori for theoretical relevance and empirical strength. For this reason, although certain emotions were omitted from our investigation, we do not believe that the results are biased, as much as potentially limited in generalizability to other emotional domains. Therefore, while assessing a broader range of relevant emotion states and emotional change is an admirable goal of future research, we believe that our investigation provided unbiased information about the nature of the relationship between eating and emotions among individuals with EDs.