Understanding birth statistics

Recently there has been a lot of hysterical press coverage about the changes in teen birth data. Let's look at one aspect of the data: statistical significance. This looks like a scary math term, but it's really not that bad.

Wikipedia's article about "statistical significance" says the following. "In statistics, a result is called statistically significant if it is unlikely to have occurred by chance. "A statistically significant difference" simply means there is statistical evidence that there is a difference; it does not mean the difference is necessarily large, important, or significant in the common meaning of the word."

For a data point to be statistically significant, it has to be just larger than the cutoff point. For example, if the cutoff point was 0.05, then a data point of 0.05001 would be statstically significant. Interestingly, the cutoff point can be chosen by the researcher. There are standard choices, but the researcher can choose any of them.

Using the interactive tool that looks at the 2006 teen birth data, the following states had changes that were not statistically significant. In other words, the birth data changes for these states is small enough that it could be random fluctuation: Idaho, Wyoming, Utah, New Mexico, Colorado, South Dakota, Nebraska, Kansas, Iowa, Wisconsin, Indiana, Tennessee, Virginia, West Virginia, Delaware, Pennsylvania, Connecticut, Massachussets, Vermont, New Hampshire, and Maine.

Another point to remember is that the data represented in the interactive tool above includes both married and unmarried teens. As I posted the other day: Table 18 in the CDC report shows percent of births to UNmarried women:
Under 15 years of age: 98.3% (a few under age 15 are married!)
Age 15: 96.8% unmarried births (3.2% married)
Age 16: 93.3% unmarried (6.7% married)
Age 17: 90.0% unmarried (10.0% married)
Age 18: 84.4% unmarried (15.6% married)
Age 19: 77.8% unmarried (22.2% married, almost 1/4)

I wonder what would happen to the birth data in the interactive tool if we looked only at births to unmarried teens. Would some states drop out of the statistically significant change?