For skewed distributions, the mean will be too small or too large depending on which tail is pulled out. This is why choice A is not the answer.
Choice B is not the answer because bimodal distributions will have the mean between the two modes, assuming the data is roughly symmetrical on either side. Though the majority of the data values is in the peaks, so its likely better to go with the mode in this case.
Choice C is wrong because A is not correct.
Choice D is the answer because the majority of the data values is either at or close to the peak point where the median and mode will also be close to as well. All three values are identical if the distribution is perfectly symmetric (eg: normal distribution).
