In election polls, do the RLC (the ratio of landline surveying to cellphone surveying), response rate, ARS (Auto Response System) influence the estimation of candidates’ approval ratings and the gap between them? Using data on 86 election polls during the 19th presidential election, this paper empirically analyzes the effects of different polling methods – RLC, response rate, and ARS – on the estimation of approval ratings of electoral candidates. The results of empirical analysis show that the approval rating of Moon Jae-in (the progressive candidate who receives overwhelming support from the 20s and the 30s) tends to increase in election polls where the proportion of cellphone surveying is high, response rate is low, and ARS is adopted. After all, the result suggests that the deep generational cleavage and the solidity of the supporters are being reflected in election polls.