Iron carbides are unmistakably associated with the active phase for Fischer-Tropsch synthesis (FTS). The formation of these carbides is highly dependent on the catalyst formulation, the activation method and the operational conditions. Because of this highly dynamic behavior, studies on active phase performance often lack the direct correlation between catalyst performance and iron carbide phase. For the above reasons, an extensive in situ Mössbauer spectroscopy study on highly dispersed Fe on carbon catalysts (Fe@C) produced through pyrolysis of a Metal Organic Framework was coupled to their FTS performance testing. The preparation of Fe@C catalysts via this MOF mediated synthesis allows control over the active phase formation and therefore provides an ideal model system to study the performance of different iron carbides. Reduction of fresh Fe@C followed by low-temperature Fischer-Tropsch (LTFT) conditions resulted in the formation of the ε′-Fe2.2C, whereas carburization of the fresh catalysts under high-temperature Fischer-Tropsch (HTFT) resulted in the formation of χ-Fe5C2. Furthermore, the different activation methods did not alter other important catalyst properties, as pre- and post-reaction transmission electron microscopy (TEM) characterization confirmed that the iron nanoparticle dispersion was preserved. The weight normalized activities (FTY) of χ-Fe5C2 and ε′-Fe2.2C are virtually identical, whilst it is found that ε′-Fe2.2C is a better hydrogenation catalyst than χ-Fe5C2. The absence of differences under subsequent HTFT experiments, where χ-Fe5C2 is the dominating phase, is a strong indication that the iron carbide phase is responsible for the differences in selectivity.