Vector Sampling

Question 1

Sensitivity

Answer 1

Probability of detecting at least one vector in a site that is actually infested (e.g. vectors can be elusive)

Question 2

Specificity

Answer 2

Probability of detecting no vectors in a site that is actually not infested (e.g. taxonomic error)

Question 3

WHO policy of reporting naive indices of vector species presence (Calculation)

Answer 3

No. of locations where vectors are found/No. of locations checked for vectors

Question 3

WHO policy of reporting naive indices of vector species presence (Calculation)

Answer 3

No. of locations where vectors are found/No. of locations checked for vectors

Question 4

True positive (Vector searches)

Answer 4

At least 1 of the searchers found a vector

Question 5

Likelihood-based methods for estimating site occupancy rates

Answer 5

Measure covariates (search round, type of ecotope, ecotope size class)
Assess relationship between imperfect detection and covariates
Produce vector-occupancy models incl full range of covariates
Assess ‘best’ model using complexity-precision trade-off (AIC)

Question 6

Types of spatial distribution

Answer 6

Random (Poisson)
Clustered
Regular

Question 7

Methods of assessing vector clusters

Answer 7

Ripley’s k-function (Point data)

Moran’s I (Area Data)

Question 8

Global cluster statistics

Answer 8

Inform whether or not vectors are significantly clustered

Question 9

Ripley’s k-function

Answer 9

Number of pairs of points separated by given distance
k-function ratios incorporating underlying population
Significance testing using MC simulation

Question 10

Moran’s I

Answer 10

Correlation coefficient for values in neighbouring units

Weighted to account for geographical proximity

Question 11

K value - Ripley’s K

Answer 11

When the observed K has a higher value than the confidence intervals surrounding the expected K, this implies there is greater than expected clustering in the kdr frequency values at that spatial scale

Question 12

Spatial scan statistic (SaTScan)

Answer 12

Point and area (centroid) data
Detects location (and time) of spatial (spatiotemporal) cluster(s)
Moving circular window – likelihood ratio test to determine if rate inside is higher/lower than rate outside the circle
For space-time cluster, uses a cylinder
Selects most likely cluster
Monte Carlo simulation to test significance
Many applications! Freeware, accounts for multiple testing, statistically robust