Abstract
The increase in mobile phone ownership is changing the sampling frame for landline telephone surveys, with a consequent impact on coverage error. This paper describes the main features of the Italian phone market - characterized by high mobile phone penetration rates - and the rising impact of mobile-phone-only (MPO) households. A survey that uses a landline sampling frame excludes MPO and no-phone households, creating a noncoverage rate of 17% in 2002. Types of phone arrangements and noncoverage vary dramatically among households: by region, household type, age, education and social class of the household. All these differences clearly introduce a non-ignorable bias in landline telephone surveys. Possible solutions are discussed from a methodological perspective. The analysis presented uses data collected in a face-to-face survey by the Italian Institute of Statistics.