Marketing Research Quiz 2

Questionnaires

structured set of questions used to colect data for research/evalutaion

"All surveys are wrong; good surveys are useful"

no survey is perfect, but a survey can be valuable if it gives useful, actionable information. Questionnaire design is more of an art than a science, and a good survey is one that helps solve the business problem, even if it is not flawless.

Sections in a survey

Request for cooperation, screening, information sought, classification data

Request for cooperation

a short intro that explains who you are, why the survey matters, and why the person should respond.

Screening question

used to determine whether the respondent fits the target population, so only qualified people continue.

Information sought

the main questions that collect the data you actually need.

Classification data

background questions such as age, gender, income, education, or other demographics, usually placed at the end.

Screening Question

used to determine whether a respondent belongs in the target population for the study. It helps identify who should continue and who should be terminated from the survey

Best flow for a typical survey

easy to sensitive & general to specific

Pretesting guidelines

Pretest the survey to check reliability, validity, question wording, instructions, survey length, fatigue, and scaling problems before full launch.

Pretesting watchouts

Watch for repeated questions, unrelated answers, too many "don't know" responses, too much end

Bad Questions

Leading, Loaded, Absolutes, Double

Sampling

the process of obtaining information from a subset of a larger group and then projecting those results to the larger group. The slides emphasize that the sample should be selected scientifically.

Sample vs Census vs Population

A population is the entire group you want information about. A census means studying the whole population. A sample is only a subset of that population. For example, all U.S. dog owners would be the population, surveying all of them would be a census, and surveying some of them would be a sample.

Population of Interest

The population of interest is the target population that meets the study's membership requirements. The slides say you must define it clearly and precisely, including who should be excluded. The more specific your target population is, the harder and more expensive it is to find the right sample.

What does N represent in statistical notation

Population size

What does n represent in statistical notation

Sample size

What is a sampling frame

the source material or list from which the sample is drawn. (geographic areas, institutions, individuals, or other units)

Sampling methods

Probability samples and nonprobability samples

Probability Samples

They give each target population element a known, non

Non

probability Samples

Convenience(Accidental) Sample

selects people because they are easy to reach and available in the right place at the right time (street interviews)

Judgement Sample

handpicks population elements based on the researcher's belief that they best serve the research purpose (choosing experts over randoms)

Snowball Sample

used for low

Quota Sample

selects respondents so the sample proportions for certain characteristics roughly match those characteristics in the target population (used to compare subgroups)

Simple Random Sample

each population element has a known and equal chance of being selected. If a sampling frame exists, you assign numbers to elements and use random numbers to choose the sample.

Systematic Sample

selects every kth element after a random start. It is sometimes called skip interval sampling. It is usually simpler, faster, and less expensive than simple random sampling.

Stratified Sample

divides the population into mutually exclusive and exhaustive groups.

Cluster (Area) Sample

divides the population into mutually exclusive and exhaustive groups and randomly selects entire groups

Level of acceptable error

as error increases, sample size decrease; as error decreases, sample size increases

What happens to sample size as the required confidence level increases

The sample size needed also increases.

What can you do if you can tolerate a lower confidence level

You can use a smaller sample.

What does being more confident usually mean in terms of sample size

It usually means a wider net unless the sample size increases.

element 1:What is the margin of error in sample size determination

It is the acceptable error (E).

element 2: What is the desired confidence level in sample size determination

It is represented by Z.

element 3: What represents variability in the population for sample size

It is represented by s for standard deviation or p for proportion.

Rule of thumb / Normal Distribution (n>=30)

For simple random samples with 30 or more observations, the sampling distribution of the mean is approximately normal.

Implications of sample sizes that are too big or too small

A sample that is too big can waste time and money and make tiny differences seem important. A sample that is too small may miss real differences and give less reliable results.

Survey error

all the stuff that makes data inaccurate Total error is made up of sampling error + non

Sampling error

the difference between sample results and the results from the whole population.

Selection error

happens when sampling is incomplete or improper.

How can sampling error be minimized

by using random selection procedures and quality control checks.

What is a noncoverage error

happens when the sampling frame is wrong or incomplete, leaving some population members out.

How can noncoverage error be minimized

by using the best possible sampling frame and checking it for accuracy and completeness.

Nonresponse error/bias

happens when chosen people do not respond, and respondents may differ from nonrespondents. It is minimized by doing everything possible to encourage response.

What is a response error

happens when people answer incorrectly, either on purpose or without realizing it.

How can response errors be minimized?

Response errors can be minimized through careful questionnaire design and understanding culture and wording effects.

What is a recording error

a mistake that occurs during data collection.

What causes recording errors

by interviewer or questionnaire bias/error.

How can recording errors be minimized

through interviewer training, careful interviewer selection, good questionnaire design, and pretesting.

What is an office error

an input, coding, processing, or analysis error that occurs after data collection.

How can office errors be minimized

by using software checks to catch illogical responses and coding mistakes.

What are incentives in research

are gifts or payments given to respondents for participating in research.

How do incentives affect participant recruitment in research

help attract participants and increase effort.

What is a potential downside of using incentives in research

Incentives may introduce bias by attracting people mainly interested in the reward.