Objective Neuropsychological test batteries are administered in person to assess cognitive function in both clinical and research settings. However, in‐person administration holds a number of logistical challenges that makes it difficult to use in large or remote populations or for multiple serial assessments over time. The purpose of this descriptive study was to determine whether a telephone‐administered neuropsychological test battery correlated well with in‐person testing. Methods Fifty English‐speaking patients without dementia, over 70 years old, and part of a cohort of patients in a prospective cohort study examining cognitive outcomes following elective surgery were enrolled in this study. Five well‐validated neuropsychological tests were administered by telephone to each participant by a trained interviewer within 2–4 weeks of the most recent in‐person interview. Tests included the Hopkins Verbal Learning Test‐Revised, Digit Span, Category Fluency, Phonemic Fluency, and Boston Naming Test. A General Cognitive Performance composite score was calculated from individual subtest scores as a Z‐score. Results Mean age was 74.9 years (SD = 4.1), 66% female, and 4% non‐White. Mean and interquartile distributions of telephone scores were similar to in‐person scores. Correlation analysis of test scores revealed significant correlations between telephone and in‐person results for each individual subtest, as well as for the overall composite score. A Bland–Altman plot revealed no bias or trends in scoring for either test administration type. Conclusions In this descriptive study, the telephone version of a neuropsychological test battery correlated well with the in‐person version and may provide a feasible supplement in clinical and research applications. Copyright © 2016 John Wiley & Sons, Ltd.