22 Introduction Telephone interviews are increasingly recognized as convenient assessment tools for epidemiological stud- ies with geographically scattered samples, especially where repeated assessments are to be made. In a pre- vious pilot study, Prince et al. (1998) described the development of a telephone-administered cognitive test battery (TACT), which aims to be responsive enough to detect small but important degrees of early decline in cognitive function. In 98 volunteer twins most of the component test scores were normally distributed with no evidence of floor or ceiling effects. Second-order factor analysis suggested that the TACT measured three dimensions of cognitive function: • logical memory; • crystallized intelligence; and • the ability to learn and apply rules to abstract mate- rial under pressure of time. Test-retest reliability coefficients were reported for a subset of 21 participants who had completed the tele- phone battery on two occasions with a two-month interval. The TACT battery was described in detail in the previous report (Prince et al., 1998). Component tests fell into two categories. First, there were tests that did not require supporting materials and lent themselves particularly readily to tele- phone administration: • The modified Telephone Interview for Cognitive Status (TICS-m) (Brandt et al., 1993). This has been validated as a sensitive and specific screening instrument for dementia. It tests orien- tation, concentration, immediate and delayed memory, naming, calculation, comprehension and reasoning. • Verbal fluency (animal naming) (Goodglass and Kaplan, 1983). Participants are asked to name in one minute as many animals as they can. International Journal of Methods in Psychiatric Research, Volume 10, Number 1 Reliability of a telephone-administered cognitive test battery (TACT) between telephone and face-to-face administration NEIL R.THOMPSON, Department of Psychology, Institute of Psychiatry, London, UK MARTIN J. PRINCE, Section of Epidemiology, Division of Psychological Medicine, Institute of Psychiatry, London, UK ALISON MACDONALD, PAK C. SHAM, Section of Genetic Epidemiology and Statistics, Division of Psychological Medicine, Institute of Psychiatry, London, UK ABSTRACT We have previously described the development and initial validation of a telephone administered cognitive test battery (TACT). This report investigates the retest reliability of the TACT battery between telephone administration and face-to-face administration and measures the concurrent validity of the test components against a standard test of cog- nitive ability, the WAIS-R UK . Data were collected on 27 participants aged 62–63 years. There was a two- to 10-month interval between time 1 (telephone) and time 2 (face-to-face) administration. Intra-class correlation coefficients for telephone versus face-to-face administration of the TACT were satisfactory for all component measures except for ‘object rotation’ and WAIS similarities. There was no evidence of systematic cheating on the telephone-administered test. Tests of concurrent validity showed stronger correlations for the TACT battery components with verbal tests rather than performance tests, as measured by the WAIS-R UK . A few administration difficulties are noted and recommendations for change are outlined. The TACT is a convenient assessment tool with potential for measuring cognitive change in epidemi- ological studies. Key words: dementia, assessment, cognitive ability, telephone interview