Accuracy is the ability to honestly measure to some decimal precision. That is, you can have an instrument that measures down to 0.0001 units and be sure that as long as the instrument is calibrated that it will perform to that level each time it is used.
Precision is another variant of the word accuracy, in my book. However, it may also be considered as resolution. In other words, an instrument might be able to read or resolve down to 0.00001 units, but it does not mean that it is accurate to that number of decimal places.
For example, you can have a volt meter that has 5 digits, but is only accurate to 4 digits.
Repeatability is the ability for an instrument to consistently return the same value when measuring something under the same environmental conditions.
Uncertainty is where a measurement gets fuzzy. It can be also expressed as a statistical term in that an instrument will return a measurement to 0.002%. It may or may not be related to the accuracy or resolution of the instrument or it could be a constraint of the measurement environment/conditions. That is, some external events may limit the accuracy or repeatability of a measurement. In which case those factors are taken into consideration and the measurement's accuracy is then degraded and expressed as such in the results.
A good example would be a poll of people. You select a representative sample of the larger universe to get a good approximation of what the statistic likelihood of some event occurring. If your universe contains 1 million entities, but you only sample 100 entities, then your degree of certainty that your results is truly representative of the whole universe is going to be less that 100%. The more samples you take the more likely that your answer is exactly correct. Many factors determine the degree of certainty in the results, so the experimenter must think carefully so as to not be confounded by unexpected or unaccounted factors.
"Almost" Good Answers: