Cell Phone-only Population and Election Forecasting in Spain: the 2012 Regional Election in Andalusia

Abstract

This article analyses the effects that excluding the Cell Phone-only Population (CPO) has on the accuracy of pre-election telephone surveys in Spain. The data were taken from the 2012 Andalusia Regional Election, where all poll-based forecasts failed. Our results show how using only landlines contributed to significant biases in voting estimates, where votes for the right-wing party (PP) were overestimated and votes for the left-wing parties (PSOE and IU) were underestimated. Moreover, and consistently with similar studies carried out in other countries, our analysis shows how under-coverage is clearly structured along important socio-demographic variables. This means that under-coverage could be causing bias in the measurements of other subjects that are important for social and political research in Spain