One study argues that it is unlikely based on the evidence available:
http://www.jstor.org/pss/2409666
Quoting part of the abstract:
More specifically, their "kamikaze" sperm hypothesis proposes that
deformed mammalian sperm are adapted to facilitate the formation and
functioning of copulatory plugs (Baker and Bellis, 1988). Here I argue
that most, maybe all, mammals are unlikely to produce nonfertilizing
sperm. First, mammals might not be able to afford to evolve
nonfertilizing sperm, given that a) fertilization is often unlikely
despite the huge numbers of sperm produced; b) production of larger
numbers of sperm is constrained, presumably because of metabolic
costs, evidence for which includes the fact that in species in which
sperm morphology and anatomy of the female reproductive tract increase
the probability of fertilization, the numbers of sperm produced is
lower than in others; and c) selection appears to act against the
production of deformed sperm. Second, some of the evidence advanced
for the existence of nonfertilizing sperm does not in fact support the
idea. Third, accessory gland secretions are sufficient on their own to
coagulate semen and produce fully functioning plugs; thus the male
that used accessory gland secretions would be at a clear advantage
over the male that diluted his fertilizing sperm with "kamikaze"
sperm; and indeed, current evidence indicates selection on accessory
glands, not sperm morphology, to enhance coagulation of semen. Fourth,
predictions made on the basis of the "kamikaze" sperm hypothesis are
not supported by quantitative comparisons of data from polyandrous and
monandrous primates (i.e., those in which several males mate with a
fertile female, and therefore in which sperm competition should be
operating, and those in which only one male mates).
