Term
|
Definition
interview - phone or in person
self administered - mail or in a group |
|
|
Term
| two general principles of questionnaire construction |
|
Definition
1. Anticipate problems
2. Relevance to respondent |
|
|
Term
|
Definition
respondent feels it is not legitimate - you might be a salesperson, have identification
***respondent feels info will be used against them - don't ask unnecessary private questions, and build categories for people to hide in.
resp. feels responses indicate lack of education or knowledge - put them at ease
resp. says time is valuable
over saturation of polls |
|
|
Term
|
Definition
relevance of the study's goals - they think that some good will come out of the poll
relevance of the questions to the study's goals.
***relevance of the questions to the respondent*** - only ask questions that apply to the resp. |
|
|
Term
| 6 Question qualities that affect responses |
|
Definition
1. Word Structure/Presentation - do not reinvent wheel, survey organizations have debugged questions to remove biases
2. ambiguity - don't use ambiguous words or ask ambiguous questions
3. formality and level or wording - no slang or acronyms and appropriate level for audience
4. no double barreled questions
5. avoid leading questions - you can lead by: tone of voice, emotional or loaded terms, NEVER cite authority (constitution)
6. Providing a middle alternative |
|
|
Term
|
Definition
allow space for writing any answer.
BENEFITS
good when all possible answers are unknown and you are trying to find the range.
allow for detail and clarification
preferable for complex issues that are not easily condensed into a few categories
useful for numerous response questions
allow for creativity and self expression |
|
|
Term
| Open Questions LIABILITIES |
|
Definition
LIABILITIES
collection or worthless info
standardization and comparison more difficult
*** requires writing skills and generally higher level of education ***
*** more time to complete***
*** requires more paper and look more menacing and thus reduce response rates*** |
|
|
Term
| Closed questions BENEFITS |
|
Definition
standardized answers into categories just for picking
BENEFITS
answers are standardized and can be compared
easy to enter data into computer saving time and money
if respondent has doubts about a question, can be explained by the available answers
easy for respondent to fill out
questionnaires are more often complete and less extraneous data is gathered
for sensitive interval questions (income) you can provide categories to hide in. |
|
|
Term
| Personal Interviews BENEFITS |
|
Definition
BENEFITS
flexibility - use filters, explain questions, use probe questions
highest response rate - gets illiterates, those with poor writing skills, and is tougher to through someone off your porch than it is to throw away a questionnaire
observation capability - assess whether resp. is lying or not
control over environment - who fills it out
question order is preserved
few questions go unanswered
in general - longer, more complex questions with more depth |
|
|
Term
| Personal Interviews LIABILITIES |
|
Definition
LIABILITIES
Most costly
social aspect may affect accuracy of responses
no opportunity for resp. to consult records
inconvenience
least anonymity
least standardization of question wording - interviewer may change wording to help resp. understand
tough on geographically dispersed population. |
|
|
Term
| Closed questions LIABILITIES |
|
Definition
LIABILITIES
respondent may not know answer and just guess
resp. frustrated if appropriate category is not listed
too many categories cannot be remembered (especially on the phone)
different interpretations of the questions goes undetected
variation may be restricted by forced choices
clerical errors - circle or check the wrong thing |
|
|
Term
| Interview as social interaction |
|
Definition
Many errors not due to social aspect:
1.Incorrectly hears the question
2.Unconscious mistakes – mean one thing say another
3.Memory errors
4.Clerical errors
5.Cheating by interviewer
Some errors are due to Social aspect:
1.Socialized to behave in prescribed ways in certain situations
2.***Don’t like the interviewer so the respondent lies
3.***Like the interviewer and you try to help by guessing what they want to hear and telling them that.
Factors that matter
Race/Ethnicity - can be dramatic across races when questions deal with race
Gender - best interviewers are middle aged women because they are the least threatening
social status -
Age - may not take young seriously |
|
|
Term
|
Definition
Cheaper in time and money
completed at resp. convenience
greater anonymity
standardized wording
no interviewer bias
can consult records
good for dispersed populations |
|
|
Term
|
Definition
Lack of flexibility - no probes, no explanations, must be straight forward
lower response rate - easily thrown away
no observation
no control over who fills it out
no control over question order
many more unanswered questions
cannot use complex questionaires |
|
|
Term
| factors affecting MAIL response rate |
|
Definition
Cover letter - stress significance of the study and individual
sponsorship of questionnaire
length
ease of filling out and mailing back
attractiveness of format - clean and professional
inducements to reply - best is worthiness of study's goals
very high rates from groups who think the study will do them some good
type of mailing - special delivery, first class etc.
when it arrives - weekend or holiday bad. |
|
|
Term
| Strengths of survey research |
|
Definition
useful for getting info on large populations and may be the only way to get that info - attitudes/Opinions
flexibility - variety of info gathered can be used in a variety of ways
info gathered can be very accurate |
|
|
Term
| Weaknesses of survey research |
|
Definition
can't avoid:
demands time and money
requires research knowledge and sophistication to do it right
difficult to disaggregate large surveys
dealing with self reports - they may lie
CAN AVOID:
Intensiveness sacrificed for extensiveness - only ask 2 questions each about a lot of topics, but no depth
most often cross-sectional as opposed to time series - usually only at one point in time. |
|
|
