Stars with excess infrared radiation from circumstellar dust are invaluable for studies of exoplanetary systems, informing our understanding of processes of planet formation and destruction alike. All-sky photometric surveys have made the identification of dusty infrared excess candidates trivial, however, samples that rely on data from Wise Infrared Survey Explorer (WISE) are plagued with source confusion, leading to high false-positive rates. Techniques to limit its contribution to WISE-selected samples have been developed, and their effectiveness is even more important as we near the end-of-life of Spitzer, the only facility capable of confirming the excess. Here, we present a Spitzer follow-up of a sample of 22 WISE-selected infrared excess candidates near the faint-end of the WISE detection limits. Eight of the 22 excesses are deemed the result of source confusion, with the remaining candidates all confirmed by the Spitzer data. We consider the efficacy of ground-based near-infrared imaging and astrometric filtering of samples to limit confusion among the sample. We find that both techniques are worthwhile for vetting candidates, but fail to identify all of the confused excesses, indicating that they cannot be used to confirm WISE-selected infrared excess candidates, but only to rule them out. This result confirms the expectation that WISE-selected infrared excess samples will always suffer from appreciable levels of contamination, and that care should be taken in their interpretation regardless of the filters applied.