In this paper, we revisit the SN1987A neutrino data to see its constraints on flavor conversion. We are motivated by the fact that most works that analyze this data consider a specific conversion mechanism, such as the MSW (Mikheyev-Smirnov-Wolfenstein) effect, although flavor conversion is still an open question in supernovae due to the presence of neutrino-neutrino interactions. In our analysis, instead of considering a specific conversion mechanism, we let the electron antineutrino survival probability $P_{\overline{e}\overline{e}}$ be a free parameter. We fit the data from Kamiokande-II, Baksan, and IMB detected spectrum with two classes of models: time-integrated and time-dependent. For the time-integrated model, it is not possible to put limits above $1\sigma$ (68% confidence level) on the survival probability. The same happens for the time-dependent model when cooling is the only mechanism of antineutrino emission. However, for models considering an accretion phase, $P_{\overline{e}\overline{e}}\sim0$ is strongly rejected, showing a preference for the existence of an accretion component in the detected antineutrino flux, and a preference for normal mass ordering when only the MSW is present.
Comment: 13 pages,10 figures. Version Accepted for publication. Some updates were made in the analysis with no significant deviations from previous results