Surveys are used to assess the needs of the community or evaluate program quality or customer satisfaction of facilities, programs or services offered.
Four key aspects of sampling
population that will be represented by the sample;
through a process of randomization;
that ensures, with a certain degree of probability;
that the sample is indeed good.
Population: the community (zip code, age, gender, activity preference etc.) of people whose opinions are sought. In most cases, this group is so large and diverse that it is impossible to ask each member of the group every question so it becomes necessary to select a sample.
Randomization: a subset of a population chosen in such a way that all individuals of the specified population have equal probability of being selected.
Probability: the probability of an occurrence of any event ranges from .00 (no chance that the event will occur) to 1.00 (event is certain to happen). Researchers are interested in determining the probability of an event.
Before a sample can be obtained the universe (all members of the group ie., all city residents, all members of the hiking club, all 6th graders, etc.) must be identified.
Population Lists and Concerns
Potential lists include: phone books, utility customers, registered voters, hospital or school records, drivers license records, etc.. Remember, not everyone has a phone, is a registered voter, or has a drivers license and therefore would be excluded from those lists.
List Concerns
Does the list contain any duplicate names? | |
Is it a complete list, any names missing? | |
Does the list contain any ineligible members? | |
Is the list accurate? Does it have correct phone and/or address information? |
Representative Sampling
After the population has been identified it is imperative that any sample represents the entire population. The table below shows the proportion of ethnic groups in the Town of Parker.
Total |
2,897 |
100 |
Any sampling done, would need to include the same proportions of each ethnic group. The same concept applies to other socio-demographic characteristics such as age, gender, income level, education, etc.. Representation also is strongly related to probability.
Probability and Sampling Error
A sample is a randomly selected subgroup of the population and therefore is not an exact representation. It is an estimate that reflects the larger population. This "estimate" is known as the "sampling error," and can be measured. The sampling error calculates the range of doubt, or degree of confidence, reflected in the sample measurement.
Two Steps of the Sampling Process
determine the kind of sample needed (population, data collection, etc.)
determine the appropriate sample selection process.
Types of Samples and Sample Selection Process
Sample Selection Factors
goals of survey | |
population being studied | |
project budget | |
project timeline | |
sample selection difficulty |
Types of Samples
simple random sample. This sample is the most commonly used in park, recreation and leisure services evaluations. It is based on the assumption that every member of a population has the same probability of being selected. The probability is the ratio of the sample size and the population size (sample size / population = random sample ratio). For example, start with a participant list (N=559) and randomly pick a starting point, say 134 then pick every 6th name (nth) in the entire list for a total sample of 93. This will make the margin of error 10%. | |
stratification. This method divides the population into groups and then uses the categorization as the starting point for the sample selection. | |
cluster sampling. This method assumes there are groups with identifiable characteristics (geographic area, ethnicity, etc.) which can be used to select the sample. | |
network sampling. A technique used to identify difficult to reach individuals or populations. This is a "snowball" method where once a few individuals are identified they are asked to identify others with similar characteristics. |
Determining Sample Size
The size of a sample will have a significant impact on the validity of the study. Primarily, sample size is determined by the degree of error that is acceptable to the researcher. Error can be calculated based on the sample size. It is usually expected to be within 3% to 5% for social sciences.
Additional Reading: Calculating Sample Size and Margin of Error
Copyright 2001. Northern Arizona University, ALL RIGHTS RESERVED