Web-Based Surveys Questions, Answers, and Designs

1
Web-Based Surveys Questions, Answers, and
Designs

• Kent L. Norman
• Department of Psychology
• Human/Computer Interaction Lab
• Laboratory for Automation Psychology
• http//lap.umd.edu

2
Reasons for Web-based Surveys

• Cost-efficiency, Dissemination-collection,
  Central control
• Automation
• Hypermedia

3
Pitfalls of Bad Design

• Good ideas don't always migrate directly to the
  Web.
• Bad design is easier to generate than good
  design.
• One bad design leads to another.

4
Interface Issues with Surveys

• Formatting and Navigation (Norman,Friedman,
  Norman Stevenson, 2000)
• Organization of Items and Navigation (Norman,
  Slaughter, Freidman, Norman, Stevenson, 2000)
• Automatic Customization (Norman, Pleskac,
  Norman, 2001
• Edits and Corrections and Navigation
• Conditional Jumps and Navigation (Norman, 2001
  Norman Pleskac, 2002)

5
Experiment

• Our experiment compared three different designs
• Manual Scrolling,
• Automatic Items
• Automatic Scrolling
• Survey included successive jumps of one, two, and
  three items.
• Respondents were 36 undergraduate students in a
  within subjects factorial design.

6
Manual Scrolling
7
Automatic Items
Automatically jumps to the next appropriate item
when alternative is clicked.
8
Automatic Scrolling
Automatically scrolls down to the next
appropriate item when alternative is clicked.
9
Results
Automatic Item was the fastest, especially for 3
follow-up questions.
10
Expanding
Clicking on a triangle expands the survey to show
follow-up or skipped questions.
11
Automatic Gray Out
Automatically grays out items to be skipped when
an alternative is clicked.
12
Guidelines

• Reduce the branching instructions to a minimum to
  reduce reading time, confusion, and perceived
  difficulty of the questionnaire.
• Automate conditional branching when possible, but
  allow the respondent to override branching if
  there is a need to do so on the part of the
  respondent.

13
Guidelines

• Hide follow-up questions to shorten the apparent
  length of the questionnaire. Make skipped
  questions available only if the respondent
  specifically wishes to view them.
• When the respondent is allowed to see and answer
  all questions, implement logic and consistency
  checks on conditional branches.

14
Guidelines

• Streamline forward movement through the
  questionnaire while allowing backtracking to view
  or change answers.
• Finally, although good design seems intuitive, it
  requires usability testing before final
  implementation.

15
Credits

• Support from the U.S. Census Bureau, Statistical
  Research Division.
• Thanks to students, Tim Pleskac, Kirk Norman,
  Nick Robb.

http//lap.umd.edu kent_norman_at_lap.umd.edu
16
References

• Norman K. L. Pleskac (January 2002).
  Conditional Branching in Computerized
  Self-Administered Questionnaires An Empirical
  Study. LAP-2002-01, HCIL-2002-02, CS-TR-4323,
  UMIACS-TR-2002-05.
• Norman K. L. (November 2001). Implementation of
  Conditional Branching in Computerized
  Self-Administered Questionnaires. LAP-2001-02,
  HCIL-2001-26, CS-TR-4319, UMIACS-TR-2002-1.
• Norman K. L., Pleskac T.J., Norman K. (May 2001)
  Navigational Issues in the Design of On-Line
  Self-Administered Questionnaires The Effect of
  Training and Familiarity. LAP-2001-01,
  HCIL-2001-09, CS-TR-4255, UMIACS-TR-2001-38.
• Norman K., Slaughter L. Friedman Z.,Norman
  K.,Stevenson R. (October 2000). Dual Navigation
  of Computerized Self-Administered Questionnaires
  and Organizational Records. LAP-2000-02,
  HCIL-2000-22, CS-TR-4192, UMIACS-TR-2000-71.
• Norman, K. L., Friedman, Z., Norman, K. D.,
  Stevenson, R. (2000). Navigational Issues in the
  Design of On-Line Self Administered
  Questionnaires (HCIL-TR-2000-10, LAP-TR-2000-01)