We tested the usefulness of name-letter preference scores as indirect indicators of self-esteem by exploring whether multiple unsupervised self-administrations of letter rating tasks within a short period of time yield useful data. We also examined whether preferences for initials and non-initial name-letters tap different aspects of self-esteem. Participants from a community sample (N = 161; 58 men and 106 women, 17-67 years, Mage = 34.57, SD = 13.28) completed daily letter rating tasks and state self-esteem questionnaires for seven consecutive days. They also completed a trait self-esteem questionnaire on the first measurement day as well as six months later. Preference scores for first-name initials were stronger but more unstable than preference scores for other name-letters. Preferences for first-name initials were primarily associated with directly measured state self-esteem whereas preferences for non-initials were primarily associated with directly measured trait self-esteem even if the latter was measured six months later. We thus showed that preferences for initials and non-initials are not simply interchangeable. Previous letter rating studies, which almost exclusively used initial preferences, should be interpreted in terms of state rather than trait self-esteem. In future studies, researchers should focus on the name-letter preference that reflects the aspect of self-esteem they wish to address.