Yesterday, we examined simple random and systematic sampling. Here are the last two types of probability sampling.
Stratified Sampling. Stratified sampling should be used when you have a number of distinct subgroups that need to be represented in the survey research. Begin by classifying the population into subgroups, such as gender, economic status, age, etc. After this you need to apply a simple random or systematic sampling method to each strata, or classified group. When using this method, you need to decide whether each subgroup should be proportionate or disproportionately represented. For example, your specific population may be split 25-75 between males and female, but proportions exist on national and global levels where the split is closer to 50-50.
Clustered Sampling. Custer sampling is usually a more practical random sampling method and is commonly used. This method is helpful when there is no list of the population, but there are defined subgroups, or clusters. Clustered sampling is done by randomly selecting subgroups and narrowing each subgroup down until you have the sample size you need. For example, you may start with random geographic subgroups and narrow it down to subgroups within those randomly selected areas. To use clustered sampling effectively, you usually need a rather large sample size.
Choosing a probability sampling method for your survey research, whether it's an interview, telephone, paper or online survey, will result in data representative of the entire population. Market researchers know the importance of selecting the survey sample to get valid results. Without valid results, all your surveying efforts are in vain. If you're conducting an employee feedback or customer satisfaction survey, choosing a random sampling method should be a no brainer. For surveys where you’re limited and a probability methods won’t work, select a non-probability sample that will yield accurate data.
Stratified Sampling. Stratified sampling should be used when you have a number of distinct subgroups that need to be represented in the survey research. Begin by classifying the population into subgroups, such as gender, economic status, age, etc. After this you need to apply a simple random or systematic sampling method to each strata, or classified group. When using this method, you need to decide whether each subgroup should be proportionate or disproportionately represented. For example, your specific population may be split 25-75 between males and female, but proportions exist on national and global levels where the split is closer to 50-50.
Clustered Sampling. Custer sampling is usually a more practical random sampling method and is commonly used. This method is helpful when there is no list of the population, but there are defined subgroups, or clusters. Clustered sampling is done by randomly selecting subgroups and narrowing each subgroup down until you have the sample size you need. For example, you may start with random geographic subgroups and narrow it down to subgroups within those randomly selected areas. To use clustered sampling effectively, you usually need a rather large sample size.
Choosing a probability sampling method for your survey research, whether it's an interview, telephone, paper or online survey, will result in data representative of the entire population. Market researchers know the importance of selecting the survey sample to get valid results. Without valid results, all your surveying efforts are in vain. If you're conducting an employee feedback or customer satisfaction survey, choosing a random sampling method should be a no brainer. For surveys where you’re limited and a probability methods won’t work, select a non-probability sample that will yield accurate data.
Powered by Compendium Blogware
Comments for More Tips for Selecting a Random Sample Survey Method