If you schedule the Gather Schema Statistics for estimated percent 10 then what it will does.
When you say "how does it differ", well, when running at 10% you will get statistics based on a sample of 10% of the data, when running at 100% you will get fully representational statistics. 10% will be quicker, because you're processing 90% less data, however the statistics won't be as accurate. If you have a lot of data skews, oddball values etc then you often run with higher sample percentages in order to ensure your sample size captures the outliers, whereas if you have very uniform data with no major swings in column values then 10% may be completely adequate.
In addition to that, You need not mention % . There is a new feature "AUTO Sampling".Its not only has accuracy close to 100%, but it also executes more quickly than when collecting at 10% using the old method.
Best Practices for Gathering Statistics with Oracle E-Business Suite (Doc ID 1586374.1)
Hope this helps!