The initial discovery that the majority of galaxies appear to be moving away from the Earth was very surprising and puzzling. It turns out that the red shift of galaxies is not the same as the Doppler effect observed in the nearby universe; it's a different type of red shift caused by the expansion of space itself. This is called a cosmological red shift. This subject generates many frequently asked questions:
Are we are at the center of the universe? No. The way most galaxies appear to be moving away has nothing to with out location. All observers in the universe will see the same thing! The distance between a galaxy and every other galaxy increases with time. A common analogy for this is to imagine yourself on a raisin in an expanding loaf of raisin dough bread. As the dough expands, the amount of dough between any two raisins also grows. if two raisins are near one another, there won't be a lot of dough between them - maybe just 2 centimeters of dough - and if the dough rises 0.25 cm per cm per hour, those two close by raisins will only be separated by 2.5 cm at the end of an hour. On the other hand, raisins on opposite ends of the loaf, say 10 cm apart, will end up 12.5 cm apart. To you, a microbe on a raisin, the nearby raisin appears to move away at 0.5 cm/hr, while the more distant raisin appears to move away at 2.5 cm per hour. Now, if you were to switch to any other raisin, you'd see the same effect. It doesn't matter what raisin you're on, all the dough is rising so all the raisins will see all the other raisins as moving away. While the bread dough does have a center, the universe doesn't, and all galaxies are moving away from each other. An observer on any planet around any star in any galaxy would measure exactly the same relation between distance and velocity that we do.
Is everything in the universe expanding? No. Every galaxy is moving away from every other galaxy, but galaxies themselves are held together by gravity. Thus, galaxies themselves are not expanding. On a scale smaller than galaxies, normal stars are not expanding either since they are also gravitationally bound. On familiar terrestrial scales, objects are held together through chemical bonds and electromagnetic forces. Just to be clear, the Earth is not expanding, nor is your hometown or your car or your head. Only objects not otherwise held together by one of the 4 primary forces are moving apart. This means that galaxies can be bound together by gravity so that they do not take part in the general expansion. The Local Group, for instance, is gravitationally bound together and we can see galaxies moving toward us on their orbits. The expansion of the universe applies only on the largest scales. Notice that the expansion is not completely smooth and uniform. There is scatter in the relation between distance and velocity caused by the gravitational interactions among galaxies and clusters of galaxies.
What was the universe like in the past? Very hot and very dense. If galaxies are moving away from each other, then they must have all been closer together in the past. If we "play the film backwards" to imagine how the universe looked in the distant past, the voids we see today must have been filled in, and galaxies must have been on average must closer together. Just as a gas will heat up when it is compressed, a universe of galaxies must heat up when it is compressed — galaxies are of course just made of an enormous number of particles. The expansion implies a time in the distant past when all the mass in the universe was concentrated in a state of extremely high temperature and density. We appear to be riding out the aftermath of an ancient and vast expansion of this dense concentration of matter.
Is the expansion of the universe an explosion? Not really. The analogy of an explosion is flawed for two reasons. First, any explosion has a center and an edge. But every galaxy is moving away from every other galaxy and an observer on any galaxy will measure the same expansion. The number of galaxies seen in any direction of the sky is similar, so we cannot use anisotropy to place ourselves within the distribution (as astronomers did with globular clusters to show that we are not at the center of the Milky Way). Put simply, astronomers looking deeper and deeper into the universe do not "run out" of galaxies in any direction. They cannot detect a center or an edge to the expansion of the universe.
Part of the reason that an explosion is a poor analogy for the expanding universe has to do with the nature of the redshift. It is natural to interpret the redshifts as Doppler shifts. The Doppler effect is familiar in many aspects of astronomy. A large planet can cause a periodic Doppler shift in the light of a nearby star. Two stars in a binary orbit show a periodic Doppler shift as seen from Earth. We learned that we can map our position in the Milky Way by measuring the Doppler shift of stars in the rotating disk. In all of these situations, astronomers measure motions with reference to a particular object or a fixed point in space. In a more familiar example, the Doppler shift of sound waves is relative to the air through which sound travels. By contrast, the modern theory of the universe holds that the red shift is caused by the expansion of space. In an explosion, the fragments fly through space, and their motion can be defined relative to the center of the explosion and the medium through which the fragments travel. In the expanding universe, galaxies are carried apart by the expansion of space itself, not by the forces of an explosion!
Are we sure galaxy red shifts are caused by cosmic expansion? The cosmic expansion hypothesis is very bold but we should try to rule out other explanations. How do we know galaxies aren't just intrinsically fast moving objects? Some researchers have argued that some galaxies have non-cosmological red shifts, based on certain curious situations in which a galaxy with a large red shift appears to be close on the sky to a galaxy at a much lower red shift. Under the cosmological assumption, they should be at very different distances. Apparent associations are intriguing, but statistically galaxies at different distances should occasionally appear coincidentally aligned on the sky, and the number of alignments observed is consistent with the expected rate of coincident. Red shift is consistently correlated with the apparent brightness and angular size of galaxies, both of which are indicators of distance, and the great bulk of evidence thus favors the cosmological interpretation.