1 / 27

Module 8 Survey sample builder

Training in monitoring and epidemiological assessment of mass drug administration for eliminating lymphatic filariasis. Module 8 Survey sample builder. Learning objectives. By the end of this module, you should understand: how to use the survey sample builder to: determine the survey design

kali
Download Presentation

Module 8 Survey sample builder

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Training in monitoring and epidemiological assessment of mass drug administration for eliminating lymphatic filariasis Module 8 Survey sample builder

  2. Learning objectives By the end of this module, you should understand: • how to use the survey sample builder to: • determine the survey design • select random clusters and children or households • the protocol for TAS

  3. Overview • The survey sample builder • Determining the survey design • Selecting randomized clusters • Selecting randomized children or households • Protocol for TAS

  4. Survey sample builder The survey sample builder is a Microsoft Excel tool that can be used to (i) automate calculations for determining the appropriate survey design and (ii) facilitate random selection of clusters and children or households from a list of randomized numbers. Evaluation unit School Community Survey design Systematic Systematic Cluster-based Census Cluster-based Census Sample size and critical cut-off Enumeration area School Random selection Household Household Child Child

  5. Preparation before sample selectionSchool-based survey • Communicate with the ministry of education to obtain a comprehensive list of all primary schools in the defined EU. • Ideally, the list of schools will be ordered according to geographical proximity rather than alphabetically. This will allow better geographical representation of the selected schools in the EU. • Obtain an average non-response rate for the schools in the EU, if available. • The non-response rate gives an estimate of the non-participant rate and should include school absenteeism, refusals and inability to collect sufficient blood for diagnostic testing.

  6. Preparation before sample selectionCommunity-based household survey • Census enumeration areas are recommended as clusters if cluster sampling is used; enumeration areas are usually the smallest areas for which census data are available. • Obtain a list of all enumeration areas in the EU. • Ideally, the enumeration areas will be listed according to geographical proximity rather than alphabetically. • Obtain enumeration area census maps. • These maps can often be obtained from the census department or bureau of statistics. A fee may be charged to obtain these maps.

  7. Determining the survey design Available for download at http://www.filariasis.us./resources.html

  8. Determining the survey design

  9. Determining the survey design

  10. Determining the survey design Survey sampling method determined by data entered into programme

  11. Selecting randomized clusters and children or households Once the appropriate survey design has been determined, the survey sample builder facilitates selection of clusters and children or households by generating a list of randomized numbers. Evaluation unit School Community Survey design Systematic Systematic Cluster-based Census Cluster-based Census Sample size and critical cut-off Enumeration area School Random selection Household Household Child Child

  12. Selecting randomized clusters Cluster sampling requires a minimum of 30 schools or enumeration areas in which children are selected. • Obtain a comprehensive list of all primary sampling units (i.e. schools, enumeration areas) in the evaluation unit. • Number them, preferably in order of geographical proximity as opposed to alphabetical order to achieve better geographical distribution of the evaluation unit. • Once all sampling units have been assigned a number, click the “Randomize” button. • Once the survey design has been determined, a list of random numbers (e.g. 10, 23, 36, 49 …) is generated. Select the schools or enumeration areas in the list according to the random numbers (i.e. 10th, 23rd, 36th, 49th …).

  13. Selecting randomized clusters The following schoolsor enumeration areas have been selected for your study: 10 23 36 49 62 76 89 102 …. A list of random numbers can be generated to facilitate the selection of clusters to be included.

  14. Selecting randomized clusters • Selection of 5–10 additional clusters is recommended as a contingency plan if the sample size falls considerably short of the target after all the clusters have been surveyed. • Additional clusters should be selected from a list of all remaining clusters, excluding those that have already been chosen. The same random selection process used for the original clusters should be used. • The survey sample builder allows selection of additional clusters.

  15. Selecting randomized children or households After selection of randomized clusters in cluster sampling or after choosing systematic sampling, children (in school-based surveys) or households (in community-based surveys) are selected randomly. • The survey sample builder calculates the appropriate sampling interval. • The survey sample builder also calculates a random starting number and the sampling interval (inverse of the sampling fraction) to generate two numbered lists (A and B) to guide the actual selection of schoolchildren or households • For school-based surveys, lists A and B are used to select the children to be tested. • For community-based surveys, lists A and B are used to select the houses in the enumeration areas to be sampled.

  16. Selecting randomized children or households In this example, the team chose list A; therefore, the first, third, fourth, fifth … house on the list or map will be selected. Every 6- and/or 7-year-old child in the house should be tested. The same list should be used throughout the survey.

  17. Example 1 • Net school enrolment ratio: 78% • Primary vector: Culex • Population of students in target grades: 18 945 • Total number of primary schools: 386 • Estimated non-response rate: 15%

  18. Example 1. Survey design

  19. Example 1. Randomized cluster selection

  20. Example 1. Randomized cluster selection

  21. Example 2 • Net school enrolment ratio: 68% • Primary vector: Anopheles • Population of students in target grades: 23 128 • Total number of enumeration areas: 284 • Estimated non-response rate: 15%

  22. Example 2. Survey design

  23. Example 2. Randomized selection of households 1 2 4 6 3 5 11 9 12 7 10 8

  24. Example 3 • Net school enrolment ratio: 95% • Primary vector: Aedes • Population of students in target grades: 2 814 • Total number of primary schools: 24 • Estimated non-response rate: 15%

  25. Example 3. Survey design Will visit every school in the evaluation unit

  26. Protocol for TAS • Define EU • Determine survey site and sampling strategy • Calculate sample size • Prepare lists • Schools (school-based survey) • Enumeration areas (community-based survey) • Test selected children • ICT (areas endemic for W. bancrofti) • BrugiaRapidTM(areas endemic for Brugia spp.) • Interpret results on basis of critical cut-off

  27. Exercise Using population data from the evaluation unit(s) defined in the ‘Evaluation unit’ module: • Use the survey sample builder to define: • the appropriate survey design • the sample size needed • the number of clusters needed for the survey • the sampling fraction • the sampling interval • Use the survey sample builder to: • select clusters to be included (if necessary) • generate lists A and B (if necessary)

More Related