208k views
1 vote
Suppose that you decide to use the capture-recapture method to estimate the size of a network. The size of the first sample (the "marked sample") is 100 nodes, the size of the second sample is 150 nodes, the intersection of the two samples is 100 nodes. Which of the following is a valid reason to be skeptical of this estimate?

A. The two samples are too small.
B. The two samples are probably not independent.
C. The variance of an estimate can never be zero.
D. The capture-recapture estimator is only applicable when the intersection of the two samples is much lower than the size of each sample.

User Woodz
by
7.6k points

1 Answer

2 votes

Final answer:

Option D: A valid reason to be skeptical of an estimate using the capture-recapture method with a marked sample of 100 nodes and a second sample of 150 nodes, where the intersection is also 100 nodes, is that the two samples may not be independent.

Step-by-step explanation:

The student's question pertains to the use of the capture-recapture method to estimate the size of a population. A valid reason to be skeptical of the estimate, given that the size of the first (marked) sample is 100 nodes, the second sample is 150 nodes, and the intersection is also 100 nodes, would be option B: The two samples are probably not independent. The independence of samples is a crucial assumption for the capture-recapture method, and an intersection equal to the size of the marked sample suggests that the recaptured individuals may not have mixed sufficiently with the unmarked population, which could distort the estimate.

It is important to consider that the capture-recapture estimator relies on the probability of recapturing marked individuals being representative of their proportion in the overall population. Inaccuracies may arise if marked individuals have a higher or lower chance of being recaptured than unmarked ones, which can happen if the individuals learn to avoid capture or seek it out (for a reward), respectively. Additionally, a significant factor that affects the reliability of the estimate is the size of the initial sample. According to statistical literature, the larger the initial sample size, the more reliable the estimate will be. However, in this case, the intersection being equal to the size of the marked sample suggests a potential flaw in the methodology or data collection process.

User Mehrdad Afshari
by
8.3k points