It depends on the characteristics of your variable. For instance, if you hav a binary variable, the distribution can be binomial or beta distribution. And if you are dealing with cost estimates, PERTH distribution and triangular are appropriate. If you are considering time to failure, your suitable distribution is weibull, exponential and geometric distribution. Finally, if you want to approximate to a true parameter, then Gaussian (normal) distribution is the best
How did you decide on the type of distribution to be applied ?
It depends on the characteristics of your variable. For instance, if you hav a binary variable, the distribution can be binomial or beta distribution. And if you are dealing with cost estimates, PERTH distribution and triangular are appropriate. If you are considering time to failure, your suitable distribution is weibull, exponential and geometric distribution. Finally, if you want to approximate to a true parameter, then Gaussian (normal) distribution is the best