Attachment D: Defining and Selecting the Survey Sample | ADMINISTRATION-5600-MANUAL

Definitions:

Survey Population: The group of people to be surveyed.
Survey Participation Criteria: All “active” consumers receiving services within a 3-month period beginning 6 months prior to the month you will conduct the survey. For example:
- The survey month is October 2011.
- Go back 6 months.
  - Month #1 = September 2011 or 201109.
  - Month #2 = August 2011 or 201108.
  - Month #3 = July 2011 or 201107.
  - Month #4 = June 2011 or 201106.
  - Month #5 = May 2011 or 201105.
  - Month #6 = April 2011 or 201104.
- The “begin date” when setting up the report in AIMS is “201104” for Month #6 and the “end date” is “201106” for Month #4.

Note #1: The data is labeled by the SFY in which the data is collected.

Note #2: In the new SFY, begin surveying services which were surveyed first in the previous SFY.

Survey Sampling Frame: The list of consumers who meet the Criteria.
Survey Sample: The group of consumers to be contacted to complete the survey.

Calculating the Necessary Sample Size & Identifying Survey Recipients

Use the Sample Size Calculator at www.raosoft.com/samplesize.html.

The number of people determined by the calculator is the Sample. The calculator inflates the Sample size to account for those who may not complete the survey.

Decide on the Population – consumers of a particular service, providers, etc.
Within the Population, extract the list(s) of consumer meeting the Criteria PSA wide and determine the total number and break down of that number by county, or by provider, or by senior center, etc.

Note #3: The AIMS HCBS report, List of Clients by Service by PSA and County will extract the consumers in individual county lists. The AIMS HCBS report, Client Labels by Service, for the same time period and service will extract the consumers' names and addresses for printing mailing labels. (AIMS.dhr.state.ga.us)

Note #4: No matter how many consumers are in the Frame, let the Sample Calculator determine the number of consumers to be contacted to complete the survey.

For example, for a 3-county PSA, the total number of consumers meeting the
Criteria PSA wide is 1,000. These consumer lists are the Frame by county.

Attachment D-1, Calculating Number and Percentages, has been provided to calculate the percentages and numbers by county or the “sets.”

By individual county Frame, calculate the percentage of consumers meeting the Criteria based on the PSA wide total of 1,000. (Enter the names of each county into Column B of Attachment D-1 and enter the number of consumers representing each in Column C.)

County #1 has 420 consumers or 42%;
County #2 has 250 consumers or 25%; and
County #3 has 330 consumes or 33%.

1^st Field on the Calculator - Enter the Confidence Interval: The Confidence Interval assists in determining how precise or the “margin of error” the sample estimate will be as calculated by the Sample Size Calculator. The smaller the percentage, the more confident the sample estimate will be.

ALWAYS ENTER 5%.

2^nd Field on the Calculator - Enter the Confidence Level: The percentage indicating how confident you can be of the true value or proportion that you are trying to measure.

ALWAYS ENTER 95%.

3^rd Field on the Calculator - Enter the total number of consumers meeting the

Criteria in the field asking for the population size. For this example, ENTER 1,000.

4^th Field on the Calculator - Enter the Estimated Response Rate/Response Distribution/*Population *Proportion Estimate: This is the proportion of the consumers receiving the survey expected to respond to a specific question in a certain way.

For the initial year of administering the survey, ENTER 50%.

For subsequent years of administering the survey, ENTER 80%.

For this example and using the Sample Size Calculator, the “Recommended Sample Size” or Sample is 278. (Enter this number into Column/Line E35 of Attachment D-1.
Apply the percentages calculated in “B” above to the Sample Size of 278.

By rounding the numbers up:

County #1 is 42% or 117 consumers;
County #2 is 25% or 70 consumers; and
County #3 is 33% or 92 consumers.
1. Identifying the Sample:
Go to http://randomizer.org/form.htm
This site will randomize multiple unique Samples from a Frame where ALL consumers receive the same mix of services. Enter the number of sets required in the 1st Field.
*In order to complete the random selection of consumers by county, renumber the consumers within each county list beginning with “1” to the largest number and add a simple alphabet code to distinguish between the counties in the PSA. (*For example: Pierce County – 1P, 2P, 3P and so on.)

Note #5: The AAA could further segment or identify any trends by provider, senior center, etc. with an additional alphabet code. (For example: Pierce County, Green Homemaker Agency – 1P-G, 2P-G, 3P-G and so on.)

For County Set #1:
- 1^st Field - Enter “1” for the number of sets you want to generate.
- 2^nd Field - Enter “117” the number per set.
- 3^rd & 4^th Fields - Number Range: Enter “1” into the “From” field and enter “420” into the “To” field
- 5^th Field - Select “Yes” to each number in a set to remain unique.
- 6^th Field - Select “Yes: Least to Greatest”
- 7^th Field - Select “Place Markers Off”
- 8^th Field - Click on “Randomize Now!”
- Print “Research Randomizer Results”
- For County Set #2:
  - 1^st Field - Enter “1” for the number of sets you want to generate.
  - 2^nd Field - Enter “70” the number per set.
  - 3^rd & 4^th Fields - Number Range: Enter “1” into the “From” field and enter “250” into the “To” field
  - 5^th Field - Select “Yes” to each number in a set to remain unique.
  - 6^th Field - Select “Yes: Least to Greatest”
  - 7^th Field - Select “Place Markers Off”
  - 8^th Field - Click on “Randomize Now!”
  - Print “Research Randomizer Results”
- For County Set #3:
  - 1^st Field - Enter “1” for the number of sets you want to generate.
  - 2^nd Field - Enter “92” the number per set.
  - 3^rd & 4^th Fields - Number Range: Enter “1” into the “From” field and enter “330” into the “To” field
  - 5^th Field - Select “Yes” to each number in a set to remain unique.
  - 6^th Field - Select “Yes: Least to Greatest”
  - 7^th Field - Select “Place Markers Off”
  - 8^th Field - Click on “Randomize Now!”
  - Print “Randomizer Results”
- For this example, mark the corresponding numbers (the numbers you assigned within each county list; not the AIMS count) identified by Randomizer.org in each respective county Frame. These are the consumers to be contacted for the survey, the Sample.