Chapter 1: Introduction to online surveys

Question 1

What is a difference between people with and without acces to the Internet?

Answer 1

People with acces are younger, richer and more educated.

Question 2

Historical events can be reflected by internet penetration. This can be seen by looking at Germany. Explain how this can be seen

Answer 2

In West Germany, penetration is 61% and in East Germany penetration is 48%

Question 3

What is quantitative research and when should it be used?

Answer 3

Research used to quantify a research problem with NUMERICAL data. Low in cost short in time. Answer a question about great groups or generalize the results

Question 4

What is qualitative research and when should it be used?

Answer 4

Research answering questions about how and why. used to gain understanding of underlying reasons, opinions or motivations. Less structured and does not require a representative sample

Question 5

What is an internet community

Answer 5

A group of people one would not easily find outside of the Internet

Question 6

What are 3 different research methodologies for qualitative research?

Answer 6

1. Passive analysis
2. active analysis
3. Web surveys

Question 7

What is passive analysis (methodology for qualitative research)

Answer 7

Analyze interactions and information on the internet. Researchers do not actively participate but passively monitor

Question 8

What is active analysis (methodology for qualitative research)

Answer 8

Researcher gets actively involved in the interactions. Often without revealing his/her true identity

Question 9

What is a Web Survey methodology for qualitative research

Answer 9

Researchers informs participants of his/her identity making it the most ethical option. People know who the researcher is.

Question 10

What are the 2 main categories of web surveys

Answer 10

1. Non-probability surveys
2. Probability web surveys
3. Difference is if sampling has been done in a (non) random manner and if everyone has an equal chance of being selected.

Question 11

What is the main downside of non-probability web surveys?

Answer 11

They are impossible to generalize to the general population due to coverage and sample errors

Question 12

What are different types of non-probability web surveys

Answer 12

1. Polls as entertainment
2. Unrestricted elf-selected surveys (recruitment via ads / high traffic sites)
3. Volunteer opt-in panels

Question 13

Name different types of probability-based web surveys

Answer 13

1. intercept survey → every nth visitor gets invited
2. List-based sample → Invite people who have acces to the internet (student surveys)
3. Web option in mixed-mode -→ web as extra alternative
4. Pre-recruited panels of internet users -→ panel members are solicited using online modes
5. Pre-recruited panels of full population -→ people with no internet or PC are accommodated

Question 14

What is the difference between volunteer panels of internet users and pre-recruited panels of internet users?

Answer 14

Pre-recruited panels use offline modes for recruitment. For example random digit dialing and address based sampling

Question 15

What are the 5 elements of the total survey error framework

Answer 15

1. Measurement error
2. Non-response error
3. Coverage-error
4. Sampling error
5. Costs

Question 16

What are the four fundamental concepts of designing a good and accurate survey?

Answer 16

1. Accurate measurements
2. High response rate
3. High coverage
4. Proper sampling

Question 17

Explain what you know about measurement error (5)

Answer 17

1. Deviation of answers retrieved and true answers
2. Could be because of survey questions, respondents and condition factors
3. Design is important in preventing this (confusing instruction, ambiguous question or poor visual design)
4. No interviewer in web surveys to guide the process
5. Does not consider environmental factors of those who are answering

Question 18

Explain what you know about non-response error (3)

Answer 18

1. Missing answers, divided in several categories
2. Non-response does not always have to cause problems
3. Only leads to error if non-respondents differ from respondents in the features that are crucial for the questionnaire

Question 19

What are the different types of non-response (3)?

Answer 19

1. Unit non-response -→ people not participating in survey
2. Partial non-response -→ individuals who drop out during survey
3. Item non-response -→ fail to provide answer to specific question(s)

Question 20

Explain what you know about coverage error (3)

Answer 20

1. Error of non-observation because of discrepancies between frame population and actual target population
2. Usually in internet vs non-internet users
3. Digital divide and internet penetration rate

Question 21

What is the Frame Population and what is the sampling frame.

Answer 21

1. Frame population = list of target population for sampling
2. Sampling frame = sample of the frame population

Question 22

What is the difference between coverage error and sampling error?

Answer 22

1. Coverage error → incorrect frame (wrong list of target population)
2. Sampling error → correct frame, incorrect sample (correct list but wrong selection)

Question 23

What are challenges and benefits of mobile surveys?

Answer 23

1. Challenges: small in size and different/difficult to operate
2. Benefits: random based sampling solution for web surveys and more freedom, potentially improving non-response rates

Question 24

In what type of research would it be preferable to ask qualitative survey questions instead of quantitative questions?

Answer 24

1. Good fit for explorative research. Answering questions like ‘how’ and ‘why’
2. Good for obtaining a profound understanding of particular groups or groups of individuals.

Question 25

Explain the difference between probability-based Web surveys and non-probability web surveys and relate this to the problem of generalizability

Answer 25

1. Non-probability means sampling is done in a non-random menner. There are people in the population with zero chance to be selected for the survey sample. Non measuring of sources of non-response
2. Probability means all members of the target population have an equal chance to be selected to participate in a study. The recruitment process is fully random. Ability to measure different types of non-response
3. Non-probability web surveys are impossible to generalize to the general population due to coverage and sample errors.

Question 26

Explain the relation between non-coverage, error, digital divide and internet penetration

Answer 26

1. Non-coverage can lead tot a greater total survey error because of a difference between frame population and target population
2. Non-coverage is caused by the fact that internet penetration is not equal and because people with and without access differ significantly.
3. THUS not taking these different group of people into account will lead to over representation of internet users thus having larger non-coverage.

Question 27

How is self-administration related to desirability bias and what does this mean for Web surveys

Answer 27

1. Self-administration is seen as less influenced by social desirability.
2. Good news for Web surveys in making this bias less of an issue

Question 28

Explain the current state and developments of internet penetration and its relation to population size

Answer 28

1. There is a wide variability in Internet penetration
2. North America and australia/europe are largest
3. Afrika and asia are smallest
4. Africa allows for rapid growth and thus their ability to participate in Web surveys
5. Population size is related to internet penetration in such a way that you’re like to pick someone from a large population, regardless if they would have internet yes or no. So example; you’re likely to pick someone from asia but that someone is likely not to have internet.