Binning 1x1 (none)
Number of pixels: 144
Number of pixels: 36
Number of pixels: 16
The advantage of binning is that there is a reduction in noise. Whenever a detector (CCD camera, film, your eye) is exposed to light, a signal, there is always a certain amount of noise associated with the image. This noise may be random or systematic and it also may come from multiple sources. A description of how much information an image has is to create a ratio between the signal and the noise in the picture (S/N).
For example every time an image is downloaded from the CCD camera, there is a bit of noise (uncertainty) in the values because the process of reading the chip and transferring the information is not perfect. This is called "read noise." It is intrinsically random and unavoidable. You can thereby imagine that a single pixel has a bit of signal and noise.
This is where the benefit of binning becomes clear. By making a larger pixel you can make a larger ratio of Signal/Noise. Imagine a box (2x2 array) of 4 pixels. Each pixel has signal and noise. If you add the values of the four pixels together you will have the ratio (4*Signal)/sqrt(4*Noise). Noise adds as the square root because it is a random process. Now, if you instead make this a single pixel it still has the same area so 4*Signal (the same amount of light) was detected. However, now you just have one instance of that read Noise. So the ratio becomes (4*Signal)/Noise. This is a larger number than the unbinned group of four pixels. Thus the same amount of light fell on the chip, but we are able to perceive more signal in the binned example. The chip is acting as if it was more sensitive!
A drawback to binning is the loss of resolution. Smaller pixels detect more portions of an object. However with sufficiently small pixels binning may not have a large impact since it is generally the sky (atmosphere/seeing) that limits the resolution and not the size of the pixels themselves which are much smaller.