Getting the finger out on random-digit dialling

To the Editor: Random-digit dialling is commonly used to select samples in telephone surveys. It has been used as a method of selecting survey samples for a diverse range of topics ranging from chronic pain to lesbian health issues.1,2 However, the method has come under criticism in recent years, primarily on the basis that a large number of residents are systematically excluded from the sample.

We have recently completed a study where we compared random- versus fixed-digit dialling in telephone surveys. Our protocol was as follows: “random-digit” dialling — a randomisation protocol was developed to identify the digit (ie, finger) to be used in dialling for each attempted telephone contact; “fixed-digit” dialling — the same digit (ie, the index finger of the right hand) was used in all contact attempts.

We found that, in both fixed- and random-digit dialling, we were repeatedly contacting the same individual. Thus, we believe that neither method is optimal in controlling for sample bias in telephone surveys. We recommend that, in future, researchers pay more attention to methods for selecting telephone numbers than to methods for selecting a finger to dial with.