The flashcards below were created by user
on FreezingBlue Flashcards.
what are the cost for different kinds of surveys?
- Face to face = $1000 per person for 1 hour
- Telephone (now $5 a minute or $300/hr)
- Mail: Paper and pencil (Dillman) about $250
- • Web - $20/mo for survey-monkey
Issues for surveys
- What you mean by a phrase or word is not what the participant understands
- People don’t tell you what they really think
What is a Q-sort
a psychological test requiring subjects to sort items relative to one another along a dimension such as "agree"/"disagree"
- Thematic Apperception Test (TAT)
Why use open ended and close ended questions?
- Open ended to make sure that you are getting their ideas not just imposing yours
- Close ended to make sure that you have usable data on items of interest to you
- Questionnaire -- no dialogue
- Interview means dialogue often voice to voice
Getting representative information
- pay attention to how you ask the questions
- do not directly ask questions
Advantages of forced choice answers
- helps minimize response bias of just wanting to look good
- useful with controversial questions
which is better for surveys, rating or ranking?
generally, ranking is better if using a single item and rating is better for scale
What types of questions should always be avoided?
double barreled questions
What are the different types of response options?
- Likert scale: never ---> always
- Semantic differential scale: Sad ---> happy. Usually deals with emotions
- Guttman scale: once a month, week, day, etc. Gets more intense as scale moves down. Used mostly for symptoms.
What types of questions should always be used with any scale?
Reversed scored questions
Generally, how many answer options should be available?
How the a survey be structured or ordered?
- general to specific or vise versa
- easy to hard to vice versa
- open ended to close ended
Two types of sampling
- Census: including everyone
- sample: Include some from a population
How to get a representative sample
- Random sampling (probability sampling)
- Much better than self-selection
response rate for
randomly selected samples
- over 40% good for mail or web.
- Over 60% for telephone;
- over 80% for face to face.
If you have
Spend them on getting a representative sample, even if it is very small
Stratified random sample
Creating layers before sampling based on characteristics (age, gender, major, etc...)
How to deal with equal size samples from unequal size populations
- have to weight the answers from the unequal populations.
- Smaller populations have less weight than larger populations if same sample size.
Other forms of sampling when cannot do random samples
- Quota sampling: Include a certain number of people who have a particular characteristic of interest
- Convenience sampling (snowball samples): Sampling from participants who are available. can get many names and randomly select from among them
one point in time
follow the same participants over time
Cross-sectional study that pulls from different groups.
People leaving your study for different reasons. (death, moved away, not interested anymore etc...)
participants from different groups dropping out at different rates
What is the difference between an open-ended question and a close-ended one?
Close-ended questions present response options while open-ended questions do not.
What is the approximate cost of one hour of telephone interviewing?
Q-sort is a technique used to reduce response bias. The participant has a stack of cardsand sorts them into piles. The number of piles might be 5, 7, 9 or 11 and the number ofcards to be sorted can vary. The piles at the ends take few cards and the piles in themiddle many.
IAT is the Implicit Association Test. It is used to reduce response bias. Participants haveto decide which words go with which other words or to decide into which columns to sortwords. Responses are measured in milliseconds. A slow response is assumed to meanthat the participant does not immediately see the association, and a fast response meansthat the participant does see the association
What is the main advantage of a forced-choice item?
They minimize response bias.
Sometimes participants favor the left side or the right side of the paper or favor apositive or negative response. What can you do to minimize the impact of such tendencies?
Reverse some of the items (and then reverse score them if you make a scale).
Define “longitudinal study.”
A longitudinal study is one that extends over time.
Even longitudinal surveys are not as good as experiments for showing causalconnections (e.g., A causes B). Why
Experiments are better than longitudinal studies because the longitudinal study does noteliminate the issue of self-selection entirely and because historical events can presentthemselves as plausible alternative hypotheses for any co-variation between the presumed cause and the presumed effect.
Define “differential attrition.”
Differential attrition occurs when different groups within a sample drop out of a longitudinal study at different rates.
When is ranking preferable to rating for a questionnaire item?
When you have a single item to measure a construct or concept.