Flashcards in Chapter 4 - Quantitative Research Deck (36):
What do surveys mostly consist of?
-Coding and analysis is simple
-Reduces variability in responses due to interviewer
Define survey method
Structured questionnaire designed to elicit specific information from respondents
Can surveys have open ended questions?
Absolutely, so something like "explain the factors that led you to buy this product" could be one
What are 4 modes of administration for surveys?
1. Telephone Interviewing
2. Personal Methods
3. Mail Methods
4. Electronic Methods
Two types of telephone interviewing?
Traditional and Computer assisted (most common)
Three types of personal survey methods?
1. In-home (interviewed at home)
2. Mall-Intercept (intercepted at mall)
3. Computer Assisted
(Lab assistant needed while participants respond to questionnaire using computer)
What are mail panels?
So for mail method where large, nationally representative sample of households that agreed to participate in periodic mail questionaires
Two type of electronic methods?
1. By e-mail
(Questions in text of email, data entered or coded)
2. Online surveys
(Dynamic and interactive. Response validation allows better data quality)
What are 4 task factors that indicate when to use personal methods?
1. Complexity of questions is high so you can clarify
2. Use of physical stimuli such as product tests where respondent needs to be there
3. Quantity of data is rich and detailed for long survey
4. Response bias, response rates are higher for more personal methods
When control of the data collection environment is needed, what kind of studies should you do?
Lab studies allow highest control, whereas internet allow lowest control
When is field force most difficult to control?
Field force is most difficult to control when personal method is used and easiest to control when internet method is used
What are the quickest and cheapest survey methods?
Online surveys are the quickest and cheapest whereas personal surveys are the slowest and most expensive
What are 4 respondent factors?
1. Perceived anonymity
(allows for more honest response)
2. Social Desirability
(Personal interaction may increase this bad thing)
3. Low incidence rate
(% of people eligible to participate in a study need to be accessible by method, so if senior, don't use online)
4. Respondent control
(When more control on when and what to answer, it is greater response rate)
What is response rate %
% of total attempted interviews that are completed
What are the response rate for the 4 methods of survey
Online: Much less!
Explain non-response bias, it'll probably be on quiz. How to decrease this?
So if you have 50% squares and circles in population, and then mostly squares respond on your survey because they maybe have more time than circles. Boom, you have response bias.
You can decrease response bias by increasing response rate. You can use a survey method with a higher response rate such as personal or phone, over mail and online.
Are behaviours and attitudes observable?
Behaviours are observable, attitudes are not observable!
To assess attitudes, we rely on what?
Define self-report measures
Respondents are asked directly for their beliefs or feelings about something
What are 3 self-report attitude scales used?
1. Graphic-Rating Scales
(visual put check along line)
2. Itemized-Rating Scales
(selecting from limited number of choices)
3. Comparative-Ratings Scale
(Subjects indicate comparison by choosing)
What are 3 types of itemized-ratings scales?
1. Summated-Ratings (Likert) Scale
(respond to statements measuring construct)
2. Semantic-Differential Scale
(Bipolar adjectives, so not at all satisfied at one end, very at other)
3. Stapel Scale
What is the comparative-Ratings scale?
When subjects indicate relative judgements or comparisons.
The constant-sum method is when you allocate a fixed sum among two or more attributes. So if you have 100 points, you allocate them to the attributes.
What is reverse scaling?
Reduces response set bias because participants realize that they have to pay more attention by reversing order of scale. so 12345, mouse stays on one side, so reverse to 54321 everynow and then.
What is Response Set Bias?
Respondents answer all questions in a similar way (towards low or high end of scale)
What are two types of measures for number of items in a scale assessing an object or phenomenon?
1. Global Measures
(provides overall assessment of an object or phenomenon. SINGLE ITEM)
2. Composite Measures
(This is when you have an assessment of an object or phenomenon with items to assess all relevant aspects (MULTIPLE ITEMS)
Items are usually questions. So multi question for an assessment of an object or phenomenon is usually composite measures
How many number of scale points is usual?
Usually between 5 and 9 scale points
What happens when you have a greater number of scale points?
Causes greater reliability
How do you reduce response error when subject is merely guessing?
Adding a don't know or not sure category, however it amy encourage respondents to just not think about an answer
What are the 10 steps of questionnaire design?
1. Specifying what information to be sought
2. Determine the mode of administration
3. Determine content of individual questions
4. Determine form of response to each question
5. Determine wording of each question
6. Determine question sequence
7. Determine physical characteristics of questionnaire
8. Develop a recruiting message
9. Reexamine steps 1-8 and revise
10. Pretest Questionnaire and revise
What are the two forms of response to each question?
1. Open-ended questions
2. Close-ended questions
What 3 criteria when making close-ended questions?
(all possible responses should be included)
2. Mutually Exclusive
(No overlapping response categories within question)
3. Response order bias
(when applicable, randomly order response categories to avoid this bias)
What are 6 things to do when determining wording of each question?
1. use simple words
2. Avoid ambiguous words or questions
3. Avoid leading questions
4. Avoid unstated alternatives
5. Avoid generalizations and estimates
6. Avoid double-barreled questions
What is question-order bias?
This is when earlier questions influence answers to later questions
What are double-barreled questions?
An example is, how satisfied are you with quality and price?
So the response would be useless because you don't know if its more about quality, or price, or what.
What are two types of recruiting messages?
1. Spoken Message
(short, effective script that increases participation rate)
2. Written Message
(Explain purpose of study, length of study, and incentive to participate)