Surveys - Research Method

Question 1

questions with discrete answer alternatives provided for the survey taker to choose from.

Answer 1

Close-ended questions

Question 2

questions where the survey taker can answer the question in their own words, without a pre-selected set of responses provided.

Answer 2

Open-ended questions

Question 3

Every individual has an equal chance of being selected from the list.

Answer 3

Simple Random sampling

Question 4

Starting at a randomly chosen point, choosing every “x” (e.g. 10th) individual from a list.

Answer 4

Systematic sampling

Question 5

Prior to sampling, the population is divided into characteristics of importance for the research, such as by gender, political party, age, etc. The population is then randomly sampled within each category or stratum at the same proportion that they exist in the full population. For example, if 45% of the overall population are Democratic party registered, then 45% of the sample is randomly selected from the Democratic party category/stratum.

Answer 5

Stratified random sampling

Question 6

Often used when stratified or simple random sampling would be difficult or expensive. Cluster elements should be as heterogeneous as possible, with each cluster a small representation of the entire population. Cluster sampling if often used for market studies or research into characteristics of geographic entities like cities, counties, regions.

Answer 6

Cluster sampling

Question 7

TYPES OF NON-PROBABILITY SURVEY SAMPLING

Answer 7

• Quota sampling - Like stratified sampling, quota sampling identifies population strata (e.g. religious affiliation), but uses a non-random sampling approach to choose the necessary number of participants per stratum (e.g. ten from each religious affiliation independent of actual population occurrence). This is often done to ensure the inclusion of a particular segment of the population, and the proportions used may or may represent the actual proportion in the population.
• Convenience sampling - The sample is taken from a group of people who are easy (convenient) to contact or reach (e.g. volunteers stopping by a booth to answer a survey). The only criteria is whether the participants agree to participate in the survey.
• Snowball sampling - Existing study subjects who are hard to find or to identify (e.g. drug users) recruit future study subjects from among their acquaintances – so that the number of study subjects grows like a rolling snowball.

Question 8

represents the amount of random “sampling error” in the survey’s results (i.e. the smaller the margin of error, the more confidence that the survey’s reported results accurately reflect the actual population being sampled). Sampling error is therefore dependent on sample size, and can be decreased by increasing the sample size.

Answer 8

Margin of Error

Question 9

combination of the sampling error and bias

Answer 9

Total Error

Question 10

Selection bias

Answer 10

• Survey undercoverage of a portion of the population
• Survey nonresponsive portion of a population (i.e. either a group unwilling or unable to respond)
• Voluntary response only surveying

Question 11

MEASUREMENT BIAS

Answer 11

• Errors occurring while recording the data – e.g. recording device or computer malfunction
• Leading questions – Often intentional bias, utilizing survey questions intended to solicit a particular response
• Inadvertent false responses given by respondents – often from the elderly who may be confused or forgetful

Question 12

Types of Surveys

Answer 12

MAILED SURVEYS

TELEPHONE & CELL PHONE SURVEYS

PERSON-TO-PERSON SURVEYS

INTERNET & SOCIAL MEDIA SURVEYS