Worked solution
Seeds: when a binomial model is an approximation
(i) In a small plot of land, there are 20 seeds being sown, of which only 4 will germinate. If 5 seeds are selected at random without replacement, find the probability that at least a seed that germinates would be selected. marks
(ii) In a huge plot of land where a large number of seeds are sown, 10% of the seeds would germinate. If 5 seeds are selected at random without replacement, find the probability that more than a seed that germinates would be selected. marks
State an assumption made in calculating the probability in (ii). mark
[2009/DHS/II/7]
CJC · Review Problems · Question 1
Answer
(i) . (ii) . Assume a constant success probability and approximately independent selections in the huge population.
(i)
Use the complement, accounting for the changing population:
.
Equivalently, . A fixed p=4/20 binomial model is not exact for this small without-replacement sample.
(ii)
Let X count germinating seeds among the five selected. For the very large population, use the source model .
.
Assumption
The source accepts either a constant germination probability of 0.1 or independence between selected seeds as the stated assumption. The binomial approximation uses both: selecting five seeds changes a huge population negligibly, so successive germination outcomes can be treated as approximately independent with constant success probability.
Common mistakes
The small-plot calculation is exact; the huge-plot binomial calculation relies on a population-size approximation.