Communication often occurs in environments where background sounds fluctuate and mask portions of the intended message. Listeners use envelope and periodicity cues to group together audible glimpses of speech and fill in missing information. When the background contains other talkers, listeners also use focused attention to select the appropriate target talker and ignore competing talkers. Whereas older adults are known to experience significantly more difficulty with these challenging tasks than younger adults, the sources of these difficulties remain unclear. In this project, three related experiments explored the effects of aging on several aspects of speech understanding in realistic listening environments. Experiments 1 and 2 determined the extent to which aging affects the benefit of envelope and periodicity cues for recognition of short glimpses of speech, phonemic restoration of missing speech segments, and/or segregation of glimpses with a competing talker. Experiment 3 investigated effects of age on the ability to focus attention on an expected voice in a two-talker environment. Twenty younger adults and 20 older adults with normal hearing participated in all three experiments and also completed a battery of cognitive measures to examine contributions from specific cognitive abilities to speech recognition. Keyword recognition and cognitive data were analyzed with an item-level logistic regression based on a generalized linear mixed model. Results indicated that older adults were poorer than younger adults at glimpsing short segments of speech but were able use envelope and periodicity cues to facilitate phonemic restoration and speech segregation. Whereas older adults performed poorer than younger adults overall, these groups did not differ in their ability to focus attention on an expected voice. Across all three experiments, older adults were poorer than younger adults at recognizing speech from a female talker both in quiet and with a competing talker. Results of cognitive tasks indicated that faster processing speed and better visual-linguistic closure were predictive of better speech understanding. Taken together these results suggest that age-related declines in speech recognition may be partially explained by difficulty grouping short glimpses of speech into a coherent message, which may be particularly difficult for older adults when the talker is female.