We examined and modeled IRAS detections of classical novae in two previous studies [Harrison & Gehrz, AJ, 96, 1001 (1988) (Paper I); 101, 587 (1991a) (Paper II)]. Here we report results of the extraction of IRAS data on 116 objects with properties similar to those of classical novae. These include dwarf novae, novalike variables, symbiotic stars, and planetary nebulae with close binary central-star systems. The detection rate for the classical novae discussed in Papers I and II was about 50%, but here we find a much higher detection rate of 80% for a similarly selected sample of related objects. In Paper II, we reported moderate success in explaining IRAS detections of novae by line emission from the ejecta. This explanation does not appear viable for detections of dwarf novae and novalike variables. We address more fully the problem of source confusion, and find that source confusion can account for many of the detections of novae at low galactic latitudes. The high detection rates of the other objects, as well as the detections of many high latitude novae, are difficult to explain. Peculiar energy distributions are reported for a sample of planetary nebulae with close binaries at their centers. The IRAS energy distributions of symbiotic stars, symbiotic planetary nebulae, and classical novae are sufficiently different that IRAS color-color plots can be used to distinguish between them. The dwarf novae and novalike variables generally fall within the area defined by the colors of classical novae in such a diagram.