Introduction - Random telephone surveys are frequently used tools in public health research. The quality of data from these surveys is assessed primarily by the quality of the raw sample structure in terms of sociodemographic variables and the level of refusal rates. In both scientific and institutional publications, these two indicators are generally presented; however, when it comes to refusal rates, authors do not provide details. Objective - The purpose of this paper is to highlight the current limitations of the general and non-standardized presentation of refusal rates (without accompanying indicators) and to propose a solution to this problem. Method - Based on two random surveys conducted according to an identical survey protocol, we will detail the method of calculating refusal rates for two surveys: the Nicolle survey conducted by INPES and the KABP AIDS survey by ORS Île-de-France. The breakdown of telephone numbers will be standardized for both surveys to make the various indicators comparable. Results - Before harmonization, the Nicolle survey reports the following refusal rates for numbers listed in the directory: household refusal rate (TRM) = 28.1%, individual refusal rate (TRI) = 4.2%. Similarly, for the KABP AIDS survey, we obtain: TRM = 4.1% and TRI = 11.9%. After harmonization and calibration against a third-party presentation (detailed in the presentation), these rates become TRMNicolle = 25.8% and TRINicolle = 3.8%; TRMKABPsida = 5.3% and TRIKABPsida = 7.8%. The gap between the indicators of the two surveys therefore remains significant. However, in the context of a random survey, non-participation—whether due to refusal, dropout, or unreachability—is problematic. By supplementing the refusal rates with other indicators that we will call "failure-to-execute rates" (= unreachable person, missed appointment, dropout during the questionnaire) and "completion rates" (= 100% - TR (household + individual) - execution failure rate), we find that these two surveys have broadly the same completion rate, namely 55.1% for the Nicolle survey and 56.2% for the KABP AIDS survey. Proposals/Conclusion - Our proposals at this stage are to reiterate the importance and urgency of standardizing the presentation of telephone number usage data in random surveys and to adopt three quality indicators: the completion rate, the refusal rate, and the execution failure rate. [author’s summary]