(Jump to Reporting Possible Nova Discoveries)
We get that question a lot at AAVSO headquarters, and it deserves a detailed answer. It's always exciting to discover a new variable star on your CCD frames, or from your nightly observations. Although discoveries of novae and supernovae often get more coverage, the discovery of less prominent variables can also have benefits for scientific research. Many thousands of variable stars are known to exist, but it's likely that there are many millions more waiting to be discovered from the collected data archives of the amateur and professional observing community. So what should one do in the happy event that you discover what you think is a new variable star?
The very first step of course is to verify that it is indeed new, that it hasn't been discovered as a variable by some other observer or survey. If your object is a field star (i.e. not in a globular cluster and not in an external galaxy), the first place to look is the AAVSO's International Variable Star Index, VSX . This is the AAVSO community's online repository of known variable stars, collected by the community from a number of different sources. The General Catalogue of Variable Stars is included, but many tens of thousands more variables are also included from a number of different sources. It's a good bet that if a variable star is known and being actively observed, it's in VSX. When you go to VSX, you should have a good astrometric solution for your CCD images, and know the position of the variable to within an arcminute at least.
As a secondary source, you should also double check several other online catalogs. Doing so might be useful if the star is present in another star catalog (such as the Guide Star Catalog, or 2MASS), but might not be known as a variable. This is especially useful for checking to see if the object is a nova or large amplitude star like a Mira. If it's present in another catalog in a similar bandpass but with a much fainter magnitude, that lends weight to the possibility that it is indeed variable. You can search any number of catalogs using VizieR . A Simbad search  may also yield a result for your object, particularly if it is variable, but Simbad is not as comprehensive as VizieR.
You should also be aware that VSX does not contain all variables known in globular clusters or external galaxies. If your variable is within a cluster or external galaxy then it may already be known as a variable. One resource for globular cluster variables is Dr. Christine Clement's catalog of Globular Cluster variables, available here . VizieR will also have some catalogs of variables in external galaxies, such as the MACHO catalog of variables .
Another good practice is to make sure it's not a solar system object! The solar system is full of asteroids, comets, and other small bodies, many of which can easily reach the brightness range available to amateur telescopes. You can search for possible solar system objects if you know the postion and the time of observation to good precision by going to the Minor Planet Center's MPChecker .
Yet another check is to make sure there are no problems with the image itself. Some image defects like cosmic ray hits should be obvious as such -- they'll often have a jagged or blocky appearance in contrast to round star images on the frame. Bright moonlight or terrestrial light sources can also introduce ghost images or glints on your images. More subtle effects can sometimes occur. One example is residual charge: if you take an image that saturates a bright star, some charge may be left over at the position of the saturated object on several subsequent frames. The result will look like a fading star on subsequent frames. You should have a good feel for what your data say, and more importantly what they don't say!
One last caveat is beware unfiltered observations! Many observers have had variable star "discoveries" turned into false alarms by making unfiltered observations with CCD cameras. Many different types of variables are red in color, and will have red magnitudes far brighter than the V or visual magnitudes reported in many catalogs. CCD camera systems have much stronger response to red light, and so unfiltered observations of red objects will make them seem much brighter when compared to a bluer field or comparison star. Be extremely careful when assessing variability based on unfiltered data; a good rule to follow is that unfiltered photometry is only useful for observing objects that aren't red (how red is somewhat debatable). If you are dealing with unfiltered observations in an unfamiliar field be very careful to search all available catalogs for red stars; your "new variable" may simply be "very red".
If you think you've discovered a nova or supernova, the first place to report it is the International Astronomical Union's Central Bureau for Astronomical Telegrams  -- follow the instructions on that page to make your report. This will ensure that your discovery will reach the largest number of people as soon as possible, and will also make it easier for other observers to confirm your discovery . You should not report novae to the AAVSO first -- doing so may significantly delay confirmation of your object, and you may not receive discovery credit for it.
If your object is confirmed as a Galactic nova or bright extragalactic supernova, it will be added to the AAVSO's database of objects in short order, but we require that it be confirmed by the IAUC as a nova or supernova  before announcing it as such. In rare circumstances we will announce objects that have been submitted to us first if they are exhibiting interesting behavior that requires rapid announcement. However, it is strongly preferred that observers begin the process of reporting novae and supernovae with CBAT first. Be forewarned that CBAT may take a few days to respond to your request, and be sure to follow their reporting instructions very closely.
To reiterate one of the cautions above, be doubly suspect of any "new" object if you are observing in an unfamiliar field with an unfiltered CCD camera. Optically faint infrared sources may appear much brighter in an unfiltered CCD image. The "nova" that you've observed unfiltered might be a Mira variable or something similar. You should still pursue your finding, but make sure you check as many catalogs as possible. It is your responsibility to do these checks first.
The AAVSO International Variable Star Index (VSX) was designed from the beginning to be an easy and straightforward tool for observers to search for known variable stars as well as to submit new variable star discoveries. You'll need to do a few things in order to submit your star to VSX:
Once you've registered for VSX and logged in, you can then go to the "Submit " page, and follow the instructions on either the "New Star Form" or "New Star Wizard" links. With these tools, you can submit information about your variable to VSX moderators. These volunteers will look over your submission and determine whether your object is (a) correctly identified, (b) is variable, and (c) is previously unknown. If so, you will be notified that your submission was approved, and your object will appear in VSX. You'll receive similar notifications if your submission is rejected, or if the moderators require more information to properly assess your submission.
Note that for those of you who are planning on writing a paper on your observations and would like to keep your data confidential, you may keep the supporting observations private, but the star's identity will be made public as soon as you submit it to VSX. We encourage you to submit your objects as soon as you discover them, but please be aware that the community will at least be aware that your variable star exists as soon as you submit it to VSX, even if they can't see your data.
All observers should be proud of the work that they do in discovering new observations, and an important part of the scientific work of variable star observing is sharing your observations with the community. One aspect of this is reporting your observations to the AAVSO, and you can do this as soon as your star is approved and entered into VSX. Another important aspect of variable star observing is publication -- sharing your work in a formal way with the astronomical community through publication in a research journal. There are a number of journals that astronomical researchers use to learn about new discoveries, and the AAVSO has its own journal for variable star research -- the Journal of the AAVSO .
If you've discovered a new variable, you can write a paper about your discovery observations either before or after you submit to VSX. Purely observational papers are acceptable, and you do not need to be an astrophysicist in order to create a paper of value to the community. For example, you could easily create a paper detailing the circumstances of your discovery, your observing procedure, your observations and their analysis, and the results of your data analysis. Often, such a paper contains all the information another researcher would need to duplicate your results, and if need be to either derive important physical information about the star or to determine what future observations are needed to obtain such physical information.
If you have not written a paper before, you should read through the materials from the AAVSO's Publication Workshop held during the 95th Spring Meeting in 2006. The presentations and other workshop materials are available here . When you submit your paper to the Journal of the AAVSO, it will be read by the editor, and if accepted for consideration it will then be sent to a referee who will make a recommendation as to whether it should be published. Again, your paper does not need to be comprehensive, but it should be as complete as you can make it given the data that you have.
If you are interested in writing a paper on your discovery, we encourage you to read recent articles in the Journal of the AAVSO. If you're not yet familiar with variable stars and their classifications, we encourage you to learn more about the subject. Two good introductory texts are Percy's "Understanding Variable Stars", and Sterken & Jaschek's "Light Curves of Variable Stars".
The AAVSO community is very diverse, and there are many people who are interested in variable stars and enthusiastic about helping new members and observers learn more about variable stars and doing variable star research. There are a number of community resources you can take advantage of if you feel you need more information: