UXPA 2013 Annual Conference Friday July 12, 2013 1:30pm - 2:30 pm ET by Malcolm Kemeny, Nicole Kerber, Kathryn Summers, Noel Alton, Megan McKeever
For the more than 40 percent of the U.S. population reading at or below an eighth grade level, searching for information online can be a surprisingly difficult task. Viewing a search page full of text-heavy results can make finding the right information nearly impossible for low literacy users. With the advent of Rich Search Results displaying plain text answers, online queries could become an easier task for this at-risk population.
2. What is Low
Literacy?
What are Rich
Results?
Findings
Agenda
Rich Search Results for Low Literacy Users Kemeny, Kerber, Summers, Alton, McKeever
Research
Overview
3. What is a Low Literacy User?
Rich Search Results for Low Literacy Users Kemeny, Kerber, Summers, Alton, McKeever
4. Corporate
What is a Low Literacy User?
• 43% of Americans are considered low literacy and
read at or below an 8th grade level.
• Literacy is defined as “using printed and written
information to function in society, to achieve one’s
goals, and to develop one’s knowledge and
potential.”
• Measured using a REALM (Rapid Estimate of
Adult Literacy in Medicine) test.
REALM Test - Side A
Rich Search Results for Low Literacy Users Kemeny, Kerber, Summers, Alton, McKeever
5. Corporate
Recruiting Low Literacy Users
• Contacted various nonprofits and literacy groups
in Baltimore with little success before settling on
ad-hoc recruiting.
• Target group is unlikely to have consistent
contact method and are difficult to schedule.
• Network recruit - participants bring their
family/friends to participate.
Rich Search Results for Low Literacy Users Kemeny, Kerber, Summers, Alton, McKeever
6. Working With Low Literacy Users
• Tread carefully.
• Ask, “Do you think someone else might have problems using this
website? What do you think they would find confusing?”
• Be respectful and friendly.
• Build a relationship.
• Be prepared with all materials to get session completed quickly.
Rich Search Results for Low Literacy Users Kemeny, Kerber, Summers, Alton, McKeever
7. Web Use by Low Literacy Populations
Rich Search Results for Low Literacy Users Kemeny, Kerber, Summers, Alton, McKeever
8. Corporate
Low Literacy Web Use
• Read every word over scanning the page.
• Become disoriented on the page easily - scrolling is
difficult.
• Accept what they see online as truth without verifying or
reading any additional content.
• Less likely to access information online.
• Easily confused on pages.
• Make decisions without the benefit of context.
Rich Search Results for Low Literacy Users Kemeny, Kerber, Summers, Alton, McKeever
9. What are Rich Results?
Rich Search Results for Low Literacy Users Kemeny, Kerber, Summers, Alton, McKeever
10. Corporate
• Information seeking aid incorporated into search engines.
• Built to provide answers to queries on the primary results
page.
• Presented above the rest of the search results in large,
visually distinct, unambiguous text.
• Increases user satisfaction by presenting relevant
content where it can easily be found.
• Relatively new. Google started presenting rich results in
early 2010.https://support.google.com/webmasters/answer/99170?hl=en
Rich Search Results for Low Literacy Users Kemeny, Kerber, Summers, Alton, McKeever
Rich Search
Results
11. Corporate
• Style of results vary by topic.
• Usually come from a “do” search or a
simple question such as, “What day is
thanksgiving?”
• 20-30% increase in click through rates for
websites that support rich results.
http://www.wpromote.com/blog/seo/smx-advanced-2012-day-1-part-1/
Rich Search Results for Low Literacy Users Kemeny, Kerber, Summers, Alton, McKeever
Rich Search
Results
12. Corporate
Rich Search
Results
• Rich results are still in their
infancy and not fully supported.
• Search engines must continue to
improve support for misspellings,
phrasing, and formatting.
Rich Search Results for Low Literacy Users Kemeny, Kerber, Summers, Alton, McKeever
13. Search Behavior Study
Rich Search Results for Low Literacy Users Kemeny, Kerber, Summers, Alton, McKeever
14. Purpose:
•Understand the presentation of rich results in a search engine results page (SERP) as it relates
to information retrieval by low literacy users.
Test Plan:
•30 low literacy participants, who read at/below an 8th grade reading level, were recruited.
•Literacy levels were determined by administering the Rapid Estimate of Adult Literacy in
Medicine (REALM) test.
•Participants were asked to complete 4 tasks using one of three search engines; Google, Bing,
and Yahoo!. Each participant used one of the three engines exclusively throughout the test. All
participants were tested with a Tobii t60 (remote) eye tracker.
•Each task was designed to increase the likelihood of seeing rich results.
•$40 cash incentive for one hour of participation.
Rich Search Results for Low Literacy Users Kemeny, Kerber, Summers, Alton, McKeever
Methodology
15. Methodology
Test Scenarios:
• You and a friend want to go see the movie Looper tonight at the Landmark theatre. What
are the showtimes for tonight?
• Your daughter has a question on her homework about the word fortuitous. You want to
make sure you give her the right answer. Look up what the word means on the web.
• You are planning your families’ thanksgiving dinner. Look up what day Thanksgiving is
this year.
• Your son has been invited to a birthday party at Port Discovery. You want to call and ask
about parking. Find the phone number for Port Discovery.
Rich Search Results for Low Literacy Users Kemeny, Kerber, Summers, Alton, McKeever
16. Participant
Demographics
Rich Search Results for Low Literacy Users Kemeny, Kerber, Summers, Alton, McKeever
• 30 Total Participants.
• 56% access the Internet primarily
through a mobile device.
• 60% rated themselves as very
comfortable using the Internet.
17. Results Overview
• When some users reached a results page with rich results, they were able to
understand, interact with, and learn from rich results
• Identified THREE key behaviors among participants which limited their ability to learn
from rich results -
1. Participants used natural language queries, which search engines had trouble
providing rich results for.
2. Participants relied heavily on autocomplete, which often led them down the wrong
search path due to misspellings.
3. Participants often overlooked rich results when they were displayed.
Rich Search Results for Low Literacy Users Kemeny, Kerber, Summers, Alton, McKeever
18. Success with Rich Results
Rich Search Results for Low Literacy Users Kemeny, Kerber, Summers, Alton, McKeever
19. Corporate
Rich Search Results for Low Literacy Users Kemeny, Kerber, Summers, Alton, McKeever
• Users who saw, understood and
interacted with the rich results were
generally pleased with the information.
• Users were visibly surprised that they
were able to find the information they
were looking for easily. Users
interacting with rich results did not click
further to verify the information, and
overall task time was reduced.
Success with Rich
Results
Top: P01 - Alvin
Bottom: P16 - Earlene
20. Rich Search Results for Low Literacy Users Kemeny, Kerber, Summers, Alton, McKeever
Success with Rich
Results
P17 - Kennis
21. Rich Search Results for Low Literacy Users Kemeny, Kerber, Summers, Alton, McKeever
Success with Rich
Results
P07 - Danielle
22. P20 - David (offset)
Rich Search Results for Low Literacy Users Kemeny, Kerber, Summers, Alton, McKeever
Success with Rich
Results
23. Rich Search Results for Low Literacy Users Kemeny, Kerber, Summers, Alton, McKeever
Success with Rich
Results
Top: P26 - Victoria
24. Rich Search Results for Low Literacy Users Kemeny, Kerber, Summers, Alton, McKeever
Success with Rich
Results
Bottom: P06 - James
25.
26. Issues with Rich Results
Rich Search Results for Low Literacy Users Kemeny, Kerber, Summers, Alton, McKeever
27. Issues with Rich
Results
Task
Found and
Reported
Overlooked
Movie Showtime 1 time 7 times
Define:
Fortuitous
4 times 5 times
What day is
Thanksgiving?
8 times 7 times
Phone number
lookup
6 times 5 times
Rich Search Results for Low Literacy Users Kemeny, Kerber, Summers, Alton, McKeever
Rich Results were displayed 45 times out of a possible 120.
Rich Results were discovered and used 19 times out of 45.
28. Difficulty with Computers
• The participants during testing that tested as very
low literacy had a great deal of difficulty with the
study tasks -
• 3 participants tested as reading below a 3rd
grade level
• 9 participants tested as reading between a 4th
and 6th grade level.
• One participant entered ‘Pour Disrovr’ while trying to
find the phone number for Port Discovery’. He later
ended up looking at the Pottery Barn site.
“It’s not really hard, I just
gotta be more specific to
actually get to where I
want to go. ”
Rich Search Results for Low Literacy Users Kemeny, Kerber, Summers, Alton, McKeever
29. Difficulty with Computers
• Some test participants were not able to understand the goals of each
task. These users had trouble during the test and many became visibly
frustrated.
• Other users were familiar with web use but not comfortable using
search tools. Once participant began every search with WWW. in the
search bar.
• Existing search tools do not adequately support the needs of users new
to the web.
Rich Search Results for Low Literacy Users Kemeny, Kerber, Summers, Alton, McKeever
31. Corporate
• User queries frequently matched the
language used by the test moderator
in the prompt.
• These natural language queries
often resulted in unsuccessful
searches.
Participant Search for Task Prompt of “...find the phone number for Port Discovery”
````
Rich Search Results for Low Literacy Users Kemeny, Kerber, Summers, Alton, McKeever
Natural Language Queries
32.
33. Corporate
RECOMMENDATIONS: Search
engines should work on
providing rich results for natural
language queries. This will
simplify information seeking for
users of all literacy levels.
Rich Search Results for Low Literacy Users Kemeny, Kerber, Summers, Alton, McKeever
Natural Language Queries
35. Autocomplete
Rich Search Results for Low Literacy Users Kemeny, Kerber, Summers, Alton, McKeever
• Users relied heavily on search engine autocomplete to help guide search.
• Autocomplete was used as a way to help users avoid typing and spelling, often a
difficult activity. Users would type one letter, look to the autocomplete, type another
letter and check again.
• Autocomplete failed as user’s queries moved further from the correct spelling.
• Some users would look at the keyboard as they typed, missing the autocomplete
suggestions altogether.
36.
37. Corporate
RECOMMENDATION:
Autocorrect should support error
correction as the user types. A
search for “Thanksgaving”
should still return an
autocomplete suggestion and
rich results for Thanksgiving.
Rich Search Results for Low Literacy Users Kemeny, Kerber, Summers, Alton, McKeever
Autocomplete
38. Corporate
RECOMMENDATION: Participant
queries that receive autocorrect
results should support rich results.
Search engines currently support
autocorrection, showing rich
results for these queries can help
with information retrieval.
Rich Search Results for Low Literacy Users Kemeny, Kerber, Summers, Alton, McKeever
Autocomplete
40. Issues with Rich
Results
Task
Found and
Reported
Overlooked
Movie Showtime 1 time 7 times
Define:
Fortuitous
4 times 5 times
What day is
Thanksgiving?
8 times 7 times
Phone number
lookup
6 times 5 times
Rich Search Results for Low Literacy Users Kemeny, Kerber, Summers, Alton, McKeever
Rich Results were displayed 45 times out of a possible 120.
Rich Results were discovered and used 19 times out of 45.
47. • Rich results can be particularly beneficial to low literacy users, who tend to settle on the
first answer they find in a search.
• There were THREE behaviors from participants which impacted their ability to use rich
results -
1. Participants used natural language queries, which search engines had trouble providing
rich results for.
2. Participants relied heavily on autocomplete, which often led them down the wrong
search path due to misspellings.
3. Participants often overlooked rich results when they were displayed.
Rich Search Results for Low Literacy Users Kemeny, Kerber, Summers, Alton, McKeever
Overall
48. Overall
Rich Search Results for Low Literacy Users Kemeny, Kerber, Summers, Alton, McKeever
1. Search engines should move to provide
rich results for natural language queries.
2. Search engines should
unobtrusively support live autocorrect
inside of the autocomplete options.
4. As rich results become more common,
users will start to expect rich results.
3. Search results displayed with autocorrect
should support rich results.