A convenient way of quoting mass estimates is via Omega, the ratio of
the density contributed by some objects to the "critical density" = 3
H^2 / 8 pi G, where H is the Hubble constant and G is the universal
constant of gravitation.  The critical density is the amount of matter
that would be just sufficient to stop the expansion of the Universe
and is 10^{-29} g/cm^3.  (Of course, portions of the Universe have a
higher density than this, e.g., you, but this is an average density.)
The visible stars in galaxies contribute about 1 percent of critical
density, i.e., Omega_stars ~ 0.01; dark halos around galaxies
contribute Omega_halos ~ 0.05; mass estimates from clusters tend to
give Omega_clus ~ 0.2 (assuming the ratio of dark matter to stars is
the same in clusters as everywhere else); and theoretical
considerations (i.e., inflation) favor Omega_total = 1.  The gap
between 0.05 and 0.2 can be explained if galaxy halos extend further
out than we can measure the rotation curves, but if Omega_total = 1 we
may require extra dark matter in intergalactic space.

It's also interesting to consider the dark matter density "locally."
Within a few hundred parsecs of the Sun, this is about 0.01 Solar
masses per cubic parsec, or about 0.3 proton masses per cm^3; that's
only about 1/10 of the density of visible matter (mostly stars);
though it's much larger than critical density because we live in a
galaxy.  However, because the stars are in a thin disk while the dark
matter is more spherical, if you take an 8 kpc radius sphere centred
on the Galaxy and passing through the Sun, roughly half the mass in
this sphere is dark matter If you consider a larger sphere, e.g., out
to the Large Magellanic Cloud at 50 kpc radius, over 80% of the mass
in it is dark matter.  This estimate was first made by Jan Oort, and
the estimate of the *total* mass density near the Sun is today termed
the Oort limit in his honor.