Knowledge about what makes us happy allows for the assisting of individuals in making more informed decisions. Self-reporting of happiness is an often-used method for acquiring data on people’s well-being. The Experience Sampling Method (ESM) is an established subjective well-being measure, though it can be burdensome for participants. Another method, the Day Reconstruction Method (DRM), was introduced in 2004 by Kahneman and colleagues (2010) to alleviate this burden. This method requires participants to reconstruct their day retrospectively, and therefore may memory biases come into play. It is therefore important to research the validity of the DRM as compared to the ESM. Previous studies have attempted to compare the two methods by having participants provide both DRM and ESM happiness ratings for the same day, and found that the DRM provides reliable estimates of individual’s happiness. In such research design, however, there may be a carryover effect from the ESM to the DRM happiness ratings, due to the fact that both reports are filed on the same day, causing the previously given ESM ratings to pollute the DRM reports. Hence, this carryover effect may overestimate the strength of association between the two types of happiness ratings. This study therefore seeks to explore the strength of association between the ESM and DRM ratings when excluding this carryover effect. To that end, 128 individuals were asked to file DRM and ESM happiness reports on alternate days for one week. The two types of ratings are compared while controlling for all other factors, and the partial correlation is found to be 0.636, which is significantly lower than found by previous comparison studies. Furthermore, several moderators are examined to explore the strength of association between the ESM and DRM, among which timing, activities, location and social setting. The adjusted means of the ESM and DRM happiness ratings differ significantly during each hour of the day, and for about 30% of the examined activities. Additionally, evidence was found for the conservatism bias, stating that affect is less intense when recalled than when actually experienced. We conclude that there is indeed evidence for the carryover effect in previous studies, leading to an overestimation of the strength of association. Although the DRM is still found to be a reliable measure of experienced well-being, one should be wary of the existing discrepancy between the ESM and DRM happiness ratings