if you have low count rates, your actual distribution can be quite different. In that case you can still give the error as the standard deviation, but claiming that this is the 68% confidence interval would be wrong! In that case the most stringent approach would be to give the probability density function for your measured value. However, this function is hard to calculate, would take a lot of space, and usually nobody cares that much.
Ok, that's good enough. I had thought the whole idea wrong, like the error was an estimate to what the real value of the investigated thing is, although apparently it should just be thought as "here's the standard deviation, do what you will with it".
It of course makes sense, since the people who read the papers know what they're doing. It does cause some confusion in the undergrad labs though. People for example think that their measurements must've been off if the real value doesn't go in the one-SD interval.
Do you mean Student's T-distribution? That's what I've thought you'd use if you have too few measurements with normal error, but they're easily described with the number of measurements. Or do you use something more exotic? I've heard that some other distributions are needed in modern physics too, but wasn't sure if that was for error handling or just part of the theory (e.g. Maxwell-Boltzmann distribution).
The Student's T-distribution still assumes sampling a normal distribution. However, there is no reason to assume that you are sampling such a distribution. This is especially the case when dealing with a small number of discrete events, where you know that the normal distribution does not apply and you need to go for a Poisson distribution or even a binomial distribution.
Ok. Would an example of e.g. Poisson distribution be observing neutrinos in those underground water pools they have for that purpose?
In a few species, workers can become fertile and start laying more worker eggs by cloning themselves rather than through sexual reproduction, so I assume those can last longer after the queen dies.
http://arstechnica.com/science/2016...n-an-abandoned-polish-nuclear-weapons-bunker/
I read this just yesterday, they recently found an ant colony sitting on top of an airvent to an abandoned soviet nuclear bunker. Each year thousands of worker ants fall in and can't get back up, this has resulted in an independent, queenless, reproductionless colony that has lasted for decades in total darkness and isolation. Pretty interesting! I wonder if they would ever form similar colonies in the wild and just wait for a lost queen to wander through and join them.
Why doesn't Venus or Mars have magnetic fields?
Short answer: Nobody knows exactly.
Longer answer: For a rocky planet to generate a magnetic field it needs to fulfill at least three conditions: It needs to rotate, it needs a (partially) liquid core, and it needs thermal convection in the core. Mars and Venus both rotate, Venus definitely has a molten core, and the core of Mars is also theorized to be molten (I do not know, whether that has been decided yet, some of the Mars rovers were supposed to give insight into this). What both seem to lack is thermal convection in the core. Both seem to have had that in the past, so the question is: Why did it stop? I haven't found a satisfying answer for that.
Last I heard about it, the models for the geodynamo are not very good even for earth so it is very hard to extrapolate to other planets, where we have much less accurate measurements. Additionally, when surveying just the solar system, we have a small sample size with no clear pattern. Pinpointing the exact cause would probably require a survey of many exoplanets, which might be far away in the future.
What I've heard as the most likely explanation for that is a lack of tidal forces to increase the effects of convection. Venus has no satellite of any consequence and is almost tidally locked to the sun. Mars is farther from the sun and the two moons are tiny compared to ours.
Possible, but Ganymede has no satellite and is tidally locked to Jupiter, yet generates a magnetic field. That explanation is not convincing enough for me.
Edit: Mercury is actually the more obvious counterexample: tidally locked with the sun and no moon, but generates a magnetic field.
Possible, but Ganymede has no satellite and is tidally locked to Jupiter, yet generates a magnetic field. That explanation is not convincing enough for me.
Edit: Mercury is actually the more obvious counterexample: tidally locked with the sun and no moon, but generates a magnetic field.