Today we find a study published which indicates that children that consume caffeine may sleep less.  Besides the fact that this is an obvious and expected finding, it is interesting to note that this is the result of parents being surveyed during routine visits to the doctor.

So, besides the fact that the sleep requirements may vary significantly over the age group studied (5-12 year olds), we are now accepting, as scientific, data on the correlation between caffeine consumption and hours of sleep based on the anecdotal evidence of parents during a doctor's visit.  

Since the study was based on a survey, then it would be difficult to correlate how much caffeine was consumed, what time of the day (relative to sleep) it was consumed and then how many hours of sleep occurred.  In addition, without any controls determining the causes for interrupting sleep, it seems like quite a stretch to arbitrarily assume that there is a detrimental effect.  More importantly, it appears that the conclusions are inconsistent with how we expect caffeine effects to manifest.  In other words, there was no indication that children had trouble falling asleep, only that they appeared to sleep less.

"The results showed children ages 5 to 7 drank an average of 52 milligrams of caffeine a day, equivalent to the amount in a can of caffeinated soda. And kids of this age slept an average of 9.46 hours a night, longer than what the Centers for Disease Control and Prevention recommends.

However, kids ages 8 to 12 took in twice as much caffeine — an average of 109 milligrams a day, equivalent to nearly three cans of caffeinated soda. And kids in this age group slept an average of 8.47 hours per night, which is below the levels recommended by the CDC."

Of course, this fails to take into account that such a variation in sleep duration is already likely simply because of the age difference in the children.  This is also a blatant appeal to authority (using the CDC) since they don't have such a specific recommendation at all.
http://science.education.nih.gov/supplements/nih3/sleep/guide/info-sleep.htm

It also seems that there is no consideration as to what specific amounts of caffeine were ingested, so this aspect is also useless (considering that a 12-oz can can have as much as a 100% variation in the caffeine content).

"The authors suggest that primary care pediatricians may be able to help by screening patients for caffeine consumption and educating parents about the potentially harmful effects of caffeine."
http://www.bmedreport.com/archives/20910

Wouldn't it be nice to know what those harmful effects actually are and if they exist.  Instead what we have is another anecdotal study that does little except repeat what every parent intuitively expected.  

It is interesting to note that the only effect mentioned is a reduction in sleep, although the precise amount isn't indicated in most of the articles and without having any other information, the conclusion is reached that this is a negative effect.  We have little or no understanding about how much sleep people actually need and what determines the variation between individuals.  There have been many studies and measurements of sleep deprivation in individuals, but little that establishes what is "normal". So unless we're dealing with a sleep deprivation issue, what would be the basis for asserting that the study demonstrated any negative effect?
http://www.sleepfoundation.org/article/how-sleep-works/how-much-sleep-do-we-really-need

While there may be circumstances where the best or only insight can be obtained through the use of surveys and self-reporting, it would seem that every scientist that purports to publish such a study should have a huge warning appearing on the front of the article indicating that no actual validation of the data occurred and as a result, everything in this paper is just as speculative as if they had made it up.

Self-reporting surveys are NOT experimental data.