It is possible that no defective pen will be selected ; 1 defective will be chosen or both pens are defective.
2.)
Defective as Success (since selecting a defective pen is the point of interest.
3.)
Since selection is done without replacement
Probability of success per selection is different for each selection ;
Number of defective = 2
Number of observations = 5
P(success on first selection) = 2/5
P(success on second selection) = 1/4
Hence, X is not well approximated by a binomial random variable.
4.) if selection is done with replacement ; then then the probability of success per selection will be the Same for each selection made. Hence, X will be well approximated by a binomial Random variable.
5.) If sampling is done without replacement, then the the hypergeometric function will be a more effective approximation.