Abstract Water temperature is key to the study of aquatic ectotherm ecology, but precise measurements of individual‐based thermal experience remain difficult to validate. The stable isotope composition of oxygen in biominerals acts as a natural thermometer due to the temperature dependence of isotopic fractionation between water and mineral phases. Coefficients of published temperature‐dependent fractionation equations, however, vary among taxa (the so‐called ‘vital effect’) without apparent consistent predictors, implying that species‐specific experimental validation may be needed before inferring temperature from biomineral oxygen isotope thermometry. Here, we describe a meta‐analysis conducted to assess the influence of biological and experimental sources of variation on the coefficients of published isotope thermometry equations. We observed that the thermal sensitivity (equation slope) was resistant to any biological or experimental factors, while the isotopic spacing between water and biomineral (equation intercept) showed consistent variation. Experimental conditions and phylogeny were the two main sources of variation in equation coefficients, where experiment approaches influenced both equation intercepts and the fit of the linear regression. Our results suggest that the use of common equation slopes and generalized taxa‐specific equation intercepts may be appropriate under some circumstances. We additionally suggest that processes related to oxygen balance and osmoregulation may influence equation intercepts, and suggest further experimental work in this area. Finally, our observations provide ground for improvement for future design and reporting of biomineral thermometry experiments.