“Geographic Variations in the Cell Phone-Only Population”
In 2005, only 7.7% of adults resided in cell phone-only households. By 2010, this figure increased to 23.9%. Thus, telephone surveys that dial only landline numbers will suffer from coverage error, error introduced into a survey when not all population elements have a chance to be selected for the sample. Many telephone surveys now incorporate cell phone numbers in their sampling scheme but, for cost and operational reasons, many surveys do not dial cell phone numbers. An understanding of the differences between adults who reside in cell phone-only (CPO) households and adults who reside in households with landlines is necessary so that appropriate weighting factors can be introduced to “correct” the bias of calling only landlines. In examining these differences, we look at variables related to geography, age, education, race/ethnicity, mobility, housing, family structure, household income, political party, and others. The percentage of CPO households is considerably lower in the Northeast and is most strongly positively related to mobility and negatively related to income and age. The analysis shows that income, age, and living in a different residence one year ago explains 56.7% of the variation in CPO by state. An examination of the residuals from this regression shows that two states were vastly over predicted: Montana and South Dakota. Eliminating these states, the regression explains 68.9%. Interestingly, the percentage of a state’s area covered by cell phone companies (a supply variable) is unrelated to the percentage of CPO households.