The elemental abundances of planet host stars can shed light on the conditions of planet forming environments. We test if individual abundances of 130 known/candidate planet hosts in APOGEE are statistically different from those of a reference doppelgänger sample. The reference set comprises objects selected with the same T
eff,
log
g
, [Fe/H], and [Mg/H] as each Kepler Object of Interest (KOI). We predict twelve individual abundances (X = C, N, O, Na, Al, Si, Ca, Ti, V, Cr, Mn, Ni) for the KOIs and their doppelgängers using a local linear model of these four parameters, training on ASPCAP abundance measurements for a sample of field stars with high-fidelity (signal-to-noise ratio > 200) APOGEE observations. We compare element prediction residuals (model–measurement) for the two samples and find them to be indistinguishable, given a high-quality sample selection. We report median intrinsic dispersions of ∼0.038 dex and ∼0.041 dex, for the KOI and doppelgänger samples, respectively, for these elements. We conclude that the individual abundances at fixed T
eff,
log
g
, [Fe/H], and [Mg/H] are unremarkable for known planet hosts. Our results establish an upper limit on the abundance precision required to uncover any chemical signatures of planet formation in planet host stars.