surveys and usability - uc berkeley school of...
TRANSCRIPT
Surveys and Usability
I214 28 Oct 2008
What Info Can Be Collected via Surveys? Facts
Characteristics of respondents Demographics, experience, employment...
Self-reported behavior This instance Generally/usually Past Anticipated (who will you vote for?)
Opinions and attitudes: Preferences, opinions, satisfaction, concerns, perceptions Their expectations of their future behavior
Knowledge What do they they know about x, y, z
Some sources of error Sample, respondents Question choice Question wording Question order Method of administration
Surveyor, if administered person to person
Inferences from the data Users’ interests in influencing results
“vote and view the results”
Response Rates
% of sample who actually participate low rates may indicate bias in
responses Whom did you miss? Why? Who chose to cooperate? Why?
How much is enough? For statistically valid sample
Babbie: 50% is adequate; 70% is very good
Web surveys tend to be 3%, 5%....
Increasing response rates
Good design Short Easy to answer immediately NOT an extra step Harder to say ‘no’ to a person Explanation/request
Explain purpose of study Don’t underestimate altruism
Why you need them Incentives
Reporting back to respondents as a way of getting response Money; entry in a sweepstakes
Follow up on non-respondents (if you can)
Ways of Administering Surveys (roughly in order of expected response rate)
In person Phone Paper, in person Email (usually with a link) Web Mail
.
Active vs passive Active: solicit respondents
Send out email (with questions, or link), letters, phone Use sampling frame to develop a sample, I.e. list
Ideal: Keep track of who responds Follow up on non-respondents if possible Compare respondents/non-respondents looking for biases
Passive – “poll” Popup box: “would you take a few minutes to help us…” Link on a site
Passive: problems may include Response rate probably unmeasurable May be difficult to compare respondents to
population as a whole Likely to be biased (systematic error)
Frequent users probably over-represented Busy people probably under-represented Disgruntled and/or happy users probably over-
represented
What to look for in web survey software/site
# of questions you can ask
# of respondents How long the survey will be available Variety of question designs/formats, or at least the designs you need Kind of data analysis they support Reports they produce for you Whether you can download data into Excel etc How long the data will be available to you Most have free/lowcost versions and premium
Web survey problems Loss of context – what exactly are you asking about,
what are they responding to? Are you reaching them at the appropriate point in their
interaction with site etc? Incomplete responses Multiple submissions
NOW (PBS) using cookies to prevent repeated voting Response rate problems
Low rate Hard to calculate
Is your sampling unit the user or the user, the interaction?
Questionnaire construction Content
Goals of study: What do you need to know? What can respondents tell you?
Conceptualization Operationalization – e.g., how exactly do you define
“household with access to internet”? Question design Question ordering Short, short, short!
Respondent characteristics include: Demographics:
age, sex, race… Income Family/living circumstances
Living conditions Where Urban/suburban/rural House, apartment…etc etc.
Experience, expertise Occupation
Know how you’ll use the data. If you don’t need to know, don’t ask
Respondent characteristics
Pew Typology Quiz Burning Man Survey
What is your gender? M, F
Respondent characteristics
Pew Typology Quiz Burning Man Survey
What is your gender? M, F
What is your age? Fill in
A.
N/A
Respondent characteristics
Pew Typology Quiz Burning Man Survey
What is your gender? M, F
What is your age? Fill in
Where do you live? Drop down - Outside US,
and list of states
Respondent characteristics
Pew Typology Quiz Burning Man Survey
What is your gender? M, F
What is your age? Fill in
Where do you live? Drop down
What is your educational background?
drop down – from currently in middle school to grad degree
Respondent characteristics
Burning Man Survey
Continuous, variable translated into intervals
Pew: Race RACE What is your race? 1 White 2 Black or African-American 3 Asian or Pacific Islander 4 Mixed race 5 Native American/American Indian 6 Other (SPECIFY) 9 Don't know/Refused
Burning Man: Race
Income: Pew Last year, that is in 2006, what was your total family income
from all sources, before taxes. 1 Less than $10,000 2 $10,000 to under $20,000 3 $20,000 to under $30,000 4 $30,000 to under $40,000 5 $40,000 to under $50,000 6 $50,000 to under $75,000 7 $75,000 to under $100,000 8 $100,000 or more 9 Don't know/Refused
Respondent characteristics
Pew Typology Quiz Burning Man Survey
What is your gender? M, F
What is your age? Fill in
Where do you live? Drop down
What is your educational background?
Respondent characteristics: specific
Opinions, Behavior, Activity
Opinions, preferences, concerns
About the site: Content, organization, architecture, interface
Ease of use Perceived needs Preferences Concerns
E.g., security
Success, satisfaction Subdivided by part of site, task, purpose…
Other requirements Suggestions
Behavior Tasks (e.g., what did you do today?) Site usage, activity
Frequency; common functions – hard to answer accurately Self-reports vs observations
Time: This event Today The last time you... The last week The last month Generally...
Pew: Activity (open-ended) We'd be interested in hearing about the most amazing thing that has ever happened while you were online.
It could be a story about how something on the Web changed your life, something interesting you discovered on the Web, or something really important or even really amusing that happened during an exchange of email.
Internet Experience: Gates Foundation
Internet experience 2
Pew – networked workers
PEW: Opinions 1. Some people say they feel overloaded with information these days, considering all the TV news shows, magazines, newspapers, and computer information services. Others say they like having so much information to choose from. How about you... do you feel overloaded, or do you like having so much information available?
Feel overloaded Like having so much information available
2. Overall, do you think that computers and technology give... people MORE control over their lives, LESS control over their lives, or don't you think it makes any difference?
More Less No difference
PEW: behavior/activity 3. About how often do you go online from home?
Several times a day About once a day 3-5 days a week 1-2 days a week Every few weeks
Less often Never
Activities: Matrix question
Matrix question
Asking for follow-up
Writing Questions: Guidelines
Writing questions involves: Deciding on topics to be addressed Ordering the topics Writing a specific SET of questions Writing the ANSWERS
Comprehensive list Unambiguous Mutually exclusive (unless you ask “check all that apply”) GOAL:
consistency x respondents Unambiguous findings (insofar as possible)
Remember People skim People answer quickly
Variations on a theme What type of website do you visit? What type(s) of websites do you visit? What type of website do you visit most often? (Check
one) What types of websites do you visit most often? (Check
no more than three) What types of websites do you visit most often? (Check
all that apply) What type(s) of websites have you visited today? What type(s) of did you visit the last time you used the
internet?
Variations on a theme (cont): Types of websites
Search (e.g., Google, Yahoo) News (e.g. Google News, newyorktimes.com) Entertainment (???) Social networking (e.g., Facebook, YouTube) Video or photo sharing (e.g., YouTube, Flickr) Corporate or organizational sites (www.berkeley.edu;
www.microsoft.com; www.apple.com) Email ??
Types of questions
Start a survey > questions > examples
Writing questions – some guidelines Clear instructions
“Choose one only” or “choose as many as apply”
Bad survey questions Barreling:
Bad: Is this site clear and useful? Good: Do you find the information on this site clearly presented? Is it useful to
you?
Loose Bundling vs. Anchoring: Questions are too general and not grounded in specific behaviors. Bad: How often do you use the internet? Good: For the following internet uses, please indicate how often…LIST: email,
online shopping, online banking, reading blogs…etc etc Assumptions
Bad: How has blogging changed your life? Good: Do you blog? If yes, how well does this statement describe you? “Blogging
has changed my life.” Answer: scale from “not at all” to “very well.”
Writing questions – some guidelines
Unambiguous questions Caution re “what would you do”
Not too many items for ranking
Not too many ratings (with neutral!) No more than 7; preferably 5 or fewer
Writing: Parallel questions
White space, attractive design
Design: attractive; white space
Avoid convoluted instructions, questions, answers
People SKIM
How interesting would each of the following displays be to you?
* Historic displays with old-fashioned and famous police artifacts from the beginning of America, through the Wild West, to the mobster era and more recent times
* Modern displays where visitors go behind the badge and experience what it's like to walk in an officers shoes today, featuring the latest crime-fighting equipment and tactics
(Another ex) Avoid convoluted instructions, questions, answers
(Another ex) Avoid convoluted instructions, questions, answers
Questions should be broadly applicable (or use a skip pattern)
Recognition rather than recall P
Please evaluate the topics covered in this workshop:
Topic 1: ____________ How useful was this topic? How would you rate the presenter? What did we NOT cover that you were hoping to learn about?
Topic 2: _______________
More guidelines Edit! Use good English syntax. Focused questions people can answer briefly.
Reporting Survey Data
Some principles Reader must know how many total respondents are
represented, and must be able to calculate from data provided any numbers they may be interested in
Reader must know which responses/respondents are included and which not
Reader needs exact wording, answer choices offered, since these affect precise answers
Don’t know/refused: with large #s, there are always some. Are they included in %ages?
Any table or graphic should still make sense separated from the report. Readers don’t always read text; some just look at the “pictures” Tables and graphics get copied and re-used
Types of Variables and Measures of Central Tendency
Type Examples Measures of Central Tendency
Nominal gender, race, college major
mode
Ordinal preferences, satisfaction, opinion; age or income in intervals
mode, median
Interval (arbitrary zero point )
Year (calendar date) temperature Celsius or Fahrenheit
mode, median, or mean
Ratio age, income, time mode, median, arithmetic mean
Ratio or Nominal?
Ratio or Nominal?
1 2 3 4
Pew – Nominal variable
Burning Man - nominal
Ordinal Variable
Ratio
Tables
Questionnaire with the data filled in
Questionnaire with the data filled in
% % % %
Tables
Pew – cross tabs
Independent variable Dependent variable
Table Summarizing Key Findings