Survey Research
Chapter 6
Freddy Noriega & Leslie Barrett
CI 5390 Spring 2014
Room #174787
USA Street Survey Very Funny
A survey is one method researchers may use to collect information relative to their study.
Surveys describe the
- opinions
- behaviors
- characteristics
Types of Surveys
Creating the Research Instrument
- Maximize information
- Minimize burden on participants
- Pattern after previously successful questionnaires/interviews
- Pilot test the questionnaire or interview
Questionnaire Basics
Guidelines for Quality Questionnaire Creation
give points of reference (ranking) vs. vague questions
ask only important questions
be simple & clear
- short questions & responses
- ask explicit questions
- emphasize words for clarity (NOT)
- avoid double negatives
- avoid two part questions
give points of reference (ranking) vs. vague questions
ask only important questions
pilot test to identify & correct weak spots
Open form vs Closed form questionnaires
- open = no response limitations
- closed = restricted set of possible responses
Example
Link to a survey received via email
Web based customer satisfaction survey from Texas State
Room #174787
Discussion
Which form of questionnaire question is more effective: open-form or closed form?
Interview Basics
- Useful when researcher needs to go deeper with information collection
- Provides richer data
- Structured Interview uses an interview protocol to guide questioning
- Pilot test interview instruments
- Sequencing and social desirability bias
- Coding responses for quantitative analysis
Discussion
Socrative vote: Which has a higher potential for bias: Interviews or questionnaires?
Room #174787
Sampling
Sampling = a systematic procedure for for choosing the group to be surveyed
Samples should reflect entire population of interestAllows for impossibly large groups to be studied and findings to be generalized
Random Samples
cluster samples
stratified random samples
Freddy's awesome information about probability and all mathy stuff
Samples of Convenience
when it's impossible to get a true random sample of the population
Missing Data
can be a big problem in sampling