This study examined how laypersons assess subclinical depression in others on the basis of information about their daily lives. For 2 days, 96 participants were tracked with the Electronically Activated Recorder, a naturalistic observation method that samples ambient sounds from participants' momentary environments. Judges rated participants' levels of depression after listening to the sampled ambient sounds. Participants' depressive symptoms were assessed with the Beck Depression Inventory. Overall, judges showed little accuracy at determining participants' levels of depressive symptoms from the ambient sounds. Exploratory analyses, however, revealed that judges were more accurate among moderately and severely depressed participants, presumably because the cues judges used to assess depression (e.g., spending time alone, not socializing, not laughing) discriminated successfully only at high levels of subclinical depression.