2. INTRODUCTION
• Neyman (1938)
• double sampling
• sample selection- X-Y
• Two-phase sampling is particularly useful when the cost for observing x is
relatively cheap compared with the cost for observing Y
3. • To formalize, two-phase sampling can be described as follows:
• [Step 1] - From the finite population, select a first-phase sample A1 of size
n1 and observe x.
• [Step 2] - Treat the first-phase sample A1 as the population and select a
second phase sample A2 of size n2.
• In this case, the selection probability for the second-phase sample is often
determined by the value of x obtained from the first-phase sample
4. • Since the second-phase sample selection probability depends on the
observed value of the first-phase sample, the sample inclusion probability
for the second phase sample is a random variable in the sense that its value
is changed as the first-phase sample changes
5. Population of X (N units)
first phase
Sample (Large) n 'units
second phase
Subsample (small) n units
6. Example :
Capture-recapture estimation
• A lake has N fish.
• Catch and mark 200 fish; then release them.
• Later catch 100 fish
• Suppose 20 of the fish in the second sample are marked
• We estimate that 20% (20/100) of the fish in the lake are marked.
8. ASSUMPTIONS
• Population is closed. (N stays the same.)
• The two samples are independent.
• Application :
• The double sampling method is designed to determine
biomass by sampling in quadrats. It can be applied to a wide variety of
vegetation types, particularly grasslands and shrublands.
• This method can reduce time for taking sample , but still fairly accurate
• Clipping and weighing vegetation is expensive and tedious
9. • With this method ocular estimates of biomass are made for a small number
of quadrats
• And vegetation on the quadrats is clipped and weighed
• For the remaining quadrats only the ocular estimates are performed
10. ADVANTAGES
• Less time consuming than the harvest method ( since estimating is faster
than the clipping )
• And more accurate than the estimation method ( since estimating
production can be subjective and cause error )
• How ever it still requires a lot of training which involves weighing
representative units of a plant
11. DISADVANTAGES
• Formulae for data analysis and sample size estimation are much more
complex than for some other method