Our research assesses the feasibility of using data fusion to combine data for an alcohol survey collected from a dual-frame random digit dialing (RDD) telephone sample, with data collected from an opt-in web panel. A benefit of RDD telephone samples is that they are based on probability sampling, yet they are challenged by decreasing response rates and increasing costs. Data collection via opt-in web -panel raises concerns regarding population representativeness, yet is a fraction of the cost of a telephone sample. A hybrid approach that fuses RDD survey data with opt- in panel data is attractive from a cost perspective. The source of the RDD data is the thirteenth iteration of the National Alcohol Survey (N13), a dual-frame telephone survey conducted on landline and cell phones. N13 covers a number of alcohol related topics, including alcohol consumption and behaviors, effects of alcohol on individual lives and the lives of others, expenditures for alcohol, alcohol-attributed and non -attributed health conditions and perceptions about alcohol, as well as related factors such as emotional well-being. We hypothesize that a hybrid approach will offer probability-based prevalence estimates for drinking behaviors (e.g. drank wine in past month) and health conditions, yet allow us to collect episodic data such as volume consumed and brands, as well as expenditures for alcohol via an abbreviated version of the N13 questionnaire administered to respondents from a web panel. Our presentation focuses on identify the linking variables to fuse the data and assess conditional independence, a critical assumption for data fusion. We then compare responses from the web data with the telephone data, controlling for the linking variables. These comparisons will increase our understanding of whether web panel responses will produce estimates comparable to the telephone given that there could be mode effects and/or differences due to population representativeness.