Search and filter research plan

Contents

• Background
• Research plan
• Results

Background

We are beginning to build position search, display, and filter functionality after initial research with bidders. We have built three main ways to find positions that correspond to our initial hypothesis: that most bidders structure the job search around geographic region.

We have three ways to start finding positions:

• an open-ended keyword search bar on top for senior-level staff with established career next steps
• a drop-down list of bureaus below for mid- and entry-level staff who view the search in terms of bureaus
• a clickable map of continents for entry-level staff who think in terms of geographic regions, not bureaus or keywords

For the search results page, we also provide a left-hand nav with filters to narrow down.

Plan for this study

We have three initial hypotheses

1. Providing a clickable map will help entry-level bidders most.
2. The left-hand menu is an appropriate place for faceted search filters.
3. Bidders want to filter primarily by:

• Bureau
• Skill code
• Grade

Goals

1. Assess relative interest in the three ways of starting a position search
2. Test whether users can find the filters
3. Improve information hierarchy of filter navigation

Questions

Assess relative value of search methods

• Is there a clear "winner" of the three methods?
• Can we correlate skill code and grade to first click patterns?

Test filter findability

• Do test participants find the filter at the right time?
• Do we see any patterns in time-to-first click and/or location of first click on the search results screen?

Improve information hierarchy of filter terms

• What terms are most important? Least?
• Are we missing any important terms?

Methods

Overview

Goal	Method
Assess relative value of search methods	First click test and questionnaire; moderated usability
Test filter findability	First click test; moderated usability
Improve information hierarchy	Card sort; moderated usability

First click

First click tests measure where testers click first when asked to complete a simple task.

• Task 1: Start a completely open search
• Task 2: Start a slightly constrained search, given one search term
• Task 3: Narrow down the search to preferred positions

Card sort

Card sorts ask testers to organize concepts on physical or digital "cards" into groups.

In this case, we are asking testers to

• Choose the three most important filter terms
• Put them in order
• Enter any other terms they might want

Questionnaire

Both methods are anonymous, but a questionnaire on both will ask testers to state current grade and skill code.

Moderated usability testing

Testers will "think aloud" as they perform the same sort of tasks as in the first click test, using a similar Invision mockup.

Recruiting

We are working from a pre-existing pool of 700 volunteers from the State Department. Partners within the State Department will send the initial recruitment email to 30 - 50 bidders for each method until we have at least 30 responses.