What do you mean by "balanced"? Here is a paper I have kept around for a few years that you may find helpful. It compares responses on a Likert scale with that of a 100 point visual analog scale.
It would be useful to understand what reason you wish to use such a scale. The usual/ conventional approach is to make use of a scale with history and a researched base.
Examples being the Juster Scale an 11 point scale of probability of doing something. Usually purchasing..... Or the likelihood scale that is most common the Likert. Which has many variants with a long history.. The benefits tend to be; you can look back and find support for the use of it (part of validity assessment process for your research; but also the issued of where if falls down will (hopefully) be well documented.
If you have a special situation.. that would be exciting to look at.
while there are many great papers on this, Wright and MacRae
(2007) might help to see the sort of issues that you might wish to consider.
the link below will help in looking at scale issues if your area is about purchase probability.
If it is of an attitudinal nature (the reason Likert is so popular), then just Google in Scholar the words review and Likert and your area of study.
If this was helpful let me know if not also let me know :)
In the scales section of my website I have several scales that use 5-point frequency format, e.g., the ICAWS or QWI:
http://shell.cas.usf.edu/~pspector/scalepage.html
For examples of 5-point agreement, go to the International Personality Item Pool website (link on my site) and go to their example scales. Here's the 50-item Big Five scale:
According to Tabachinik and Fidell's book 'using multivariate statistiscs' a 7-point or more points Likert scale can be managed like a continuous scale which is always an advantage.
A single reference does not make it the "law of the land". Various disciplines approach the analysis of Likert scales differently, and you'll rarely find total agreement across disciplines. I don't know which discipline the original poster is in, but it may be helpful to ask colleagues in his discipline what the expected norm is.
In fact, there have been numerous empirical studies on what happens when you use a 5-point scale versus a 4-point scale etc., and the Krosnick & Presser chapter that I linked to above reviews this literature from a practical standpoint.