The current study aimed to evaluate the reliability of a simplified matrix sentence test in assessing older hearing-impaired adults who may also have some mild cognitive impairment in their perception of speech in noise. The simplified matrix sentence test, known as the simplified University of Canterbury Auditory Matrix Sentence Test-Paediatric (UCAMST-P) is a shorter and less cognitively demanding test that was originally designed to test paediatric listeners. We hypothesise that the shorter and simpler format of the test may make it more suitable for testing older hearing-impaired subjects that may also have cognitive impairment. A large cohort of 64 adults that were recruited through the University of Canterbury speech and hearing clinic were tested with the simplified UCAMST to determining their speech recognition threshold (SRT) in noise. Additionally, their speech recognition in noise was assessed with a clinically available test, the Quick speech in noise (QuickSIN™) test for comparison studies. Participants pure tone hearing thresholds and cognitive status were also assessed. Results showed that the simplified UCAMST gave consistent and reliable results over two trials. When different response formats were used in the simplified UCAMST, open-set and closed-set, it was found that the open-set was more reliable and had a smaller measurement error compared to the closed-set response condition. However, the participants attained better SRTs in the closed-set condition over the open-set condition. It is hypothesised this is due the cognitive advantage of being able to see the word matrix in the closed-set condition which is less cognitively demanding and offers visual cues to the possible word presented when speech intelligibility may not be optimal. Overall, it was found that the two different response conditions could not be used interchangeably. The reliability of the simplified UCAMST was found to be robust against the effect of increasing hearing loss, age and cognitive status of the listener in the closed-set condition. The largest effect found was the effect of older age and mild cognitive impairment on the measurement error in the open-set response condition. The results from the simplified UCAMST positively correlated with the results from the currently clinically used QuickSIN™ test proving the validity of the simplified UCAMST. Moreover, the SRTs derived from the simplified UCAMST had a lower measurement error compared to those found with the QuickSIN™. This study provides important evidence that the simplified UCASMT is a valid, reliable and accurate test for assessing the performance in speech in noise of older adults with a hearing impairment.