Goethite (α-FeOOH) contains two structurally-distinct oxygen sites that can potentially be independently interrogated for isotopic composition by analyzing bulk goethite δ18O (δ18Obulk) and the δ18O of water extracted from it (δ18Oextracted). The isotopic contrast between these two components may form the basis of a paleothermometer (Miller et al., 2020). We sought to better understand the behavior of goethite undergoing vacuum dehydroxylation, the key step in this potential paleothermometric technique. A refined method for making these measurements applied to a suite of natural and synthetic goethites revealed that δ18Oextracted is extremely sensitive to extraction temperature and to details of the vacuum environment in which the goethite is dehydroxylated. This sensitivity likely arises from the fundamental mechanism by which evolved water is removed from goethite in vacuum, via nm-wide pores penetrating nascent hematite surrounding dehydroxylating goethite. These pores may provide favorable conditions for back reaction that scrambles the oxygen isotopic composition, and also for kinetic mass fractionation. The character of these pores is known to vary with goethite composition, crystallinity, and extraction temperature. We identified additional sample-specific factors that affect δ18Oextracted. For example, some synthetic goethites undergo dehydroxylation at surprisingly low temperature (<90 °C) in vacuum, under conditions usually used to remove nonstoichiometric water. In addition, some of the synthetic goethites have spectroscopic characteristics suggesting a range of OH bonding environments that complicate the idea that goethite has just two structurally (and potentially isotopically) distinct oxygen sites. Instead, in fine grained or poorly crystalline goethite there may be a continuum of physisorbed water, chemisorbed water, and stoichiometric water that is simultaneously released during vacuum extraction. After investigating a wide range of extraction conditions, we determined that rapid heating to 255 ± 5 °C is optimal for achieving high degrees of water extraction with the least apparent back reaction. When applied to synthetic goethites grown between 5 °C and 70 °C, we found no compelling correlation between synthesis temperature and the contrast between δ18Obulk and δ18Oextracted. Analyses of two natural samples, known to have grown at ∼5 °C and ∼ 25 °C, revealed no significant difference in measured isotopic contrast. Together these observations suggest that goethite has no resolvable formation temperature information encoded in the bulk and extracted water oxygen isotope ratios as interrogated by the method we explored. If there is temperature-dependent isotopic contrast between the two oxygen sites in goethite, the use of the dehydroxylation method appears unable to reliably reveal it.