Improving The Safety of Homeless Young People With Mobile Phones: Values, Form and Function

Download as pdf or txt
Download as pdf or txt
You are on page 1of 10

CHI 2011 • Session: Homeless Users May 7–12, 2011 • Vancouver, BC, Canada

Improving the Safety of Homeless Young People with


Mobile Phones: Values, Form and Function
Jill Palzkill Woelfer†, Amy Iverson†, David G. Hendry†, Batya Friedman†, and Brian T. Gill‡
† ‡
The Information School Dept. of Mathematics
University of Washington Seattle Pacific University
Seattle, WA 98195-2840 Seattle, WA 98119-1997
{woelfj, ivera49, dhendry, batya}@uw.edu bgill@spu.edu

ABSTRACT ready-to-hand, if a person encounters a potentially unsafe


By their pervasiveness and by being worn on our bodies, place or situation, or happens across an accident, the mobile
mobile phones seem to have become intrinsic to safety. To phone can become an instrument for improving safety. At
examine this proposition, 43 participants, from four the same time, in some situations, the use of mobile phones
stakeholder groups (homeless young people, service may occasion moments of vulnerability and an over-
providers, police officers, and community members), were reliance on its safety functions may undermine a person’s
asked to consider how homeless young people could use overall resilience, especially if the phone malfunctions.
mobile phones to keep safe. Participants were asked to In situated use, moreover, the mobile phone can be seen as
express their knowledge for place-based safety and to ambiguous, in purpose and in form. The computational and
envision how mobile phones might be used to improve representation capacities of mobile phones are multi-
safety. Detailed analysis of the resulting data, which faceted, suitable for pleasure, for commerce, among other
included value sketches, written value scenarios, and semi- purposes. In addition, the form of the mobile phone can
structured discussion, led to specific design opportunities, also be transformed to suggest varying purposes. Related to
related to values (e.g., supporting trust and desire to help safety, this ambiguity can be seen in technological
others), function (e.g., documenting harms for future adaptations. For example, in one direction, when on-board
purposes), and form (e.g., leveraging social expectations for instruments detect poor driving or an accident, the car can
how mobile phones can be used to influence behavior). become a mobile phone with communication functions; in
Together, these findings bound a design space for how the reverse, the physical form of a mobile phone can
mobile phones can be used to manage unsafe situations. cunningly conceal a real gun. In a different vein, when the
Author Keywords mobile phone is held in particular ways it can be perceived
Safety, homeless young people, mobile phones, security, by others as particular things, such as a handgun.
Value Sensitive Design, value scenarios, value sketches How might mobile phones improve safety? Homeless
ACM Classification Keywords young people, living in urban settings, present a testing case
K.4.2 Social Issues: Miscellaneous for addressing this question. One reason is that homeless
youth are maturing and when they share information about
General Terms their whereabouts and activities, they may benefit from the
Design experience of caring others, like teens in stable families.
INTRODUCTION For a second, homeless youth face risky situations in day-
Safety is a basic human need. Not only do people wish to to-day life on the street. Specific technological solutions
keep themselves and their families safe, one of the most would not only be of great benefit to homeless youth, they
human of acts is to come to the assistance of another in a would, in general, more fully define the design space of
moment of threat or accident. mobile phones and safety for all urban dwellers.
Across societies and socio-economic classes, the mobile In this on-going research [24,25,26,27,28], therefore, we
phone is fast becoming intrinsic to safety. Not only are investigate how homeless young people manage unsafe
mobile phones pervasive, they are, like eye-glasses, situations. Specifically, in Seattle (the research site), we
generally worn on human bodies or kept nearby. By being engaged homeless young people and other stakeholders and
asked: How do homeless young people perceive safety in
Permission to make digital or hard copies of all or part of this work for their daily lives in urban places and how do they bring
personal or classroom use is granted without fee provided that copies are mobile phones into these perceptions? What kinds of unsafe
not made or distributed for profit or commercial advantage and that situations do they encounter? What stakeholders and
copies bear this notice and the full citation on the first page. To copy
otherwise, or republish, to post on servers or to redistribute to lists, institutions are implicated, and how do mobile phones
requires prior specific permission and/or a fee. interpenetrate? How do homeless youth avoid or mitigate
CHI 2011, May 7–12, 2011, Vancouver, BC, Canada. unsafe situations, and when they do occur, how are they
Copyright 2011 ACM 978-1-4503-0267-8/11/05....$10.00.

1707
CHI 2011 • Session: Homeless Users May 7–12, 2011 • Vancouver, BC, Canada

responded to? Finally, how do and how could the around, society’s dominant norms. People without shelter
informational, communication, and form-based features of become without “place.” In turn, by their efforts to live,
mobile phones be used to help keep young people safe? homeless people can be seen as threatening to both the
To address these questions we engaged multiple physical and psychological well-being of others [6,14].
stakeholder groups, as well as created a strong link between Responding to citizens’ feelings of unease, city
empirical research of current conditions and generative co- governments, including Seattle, often pass civility laws.
design activities that envision future possibilities. In this These laws, in brief, aim to improve safety by prohibiting
paper, though, we focus primarily on the empirical findings. such public behaviors as sitting on sidewalks, sleeping in
parks, and loitering in areas of drug-dealing. However,
BACKGROUND
homeless people, in part, by their living circumstances and
HCI Work Related to Homelessness and to Teen Safety
by their search for dignity often violate such prohibitions
Recently the field of HCI has begun to address questions
and thereby risk incarceration [22]. In fact, field work has
about how homeless people adopt and make use of digital
shown that civility laws tend to be disproportionately
and mobile technology and how specific applications can be
enforced against people of lower socioeconomic standing,
designed to improve their welfare. One major finding of
particularly people who are homeless [1].
this work is that homeless people, like most people in
developed and developing nations, desire digital technology The linkages between civility laws, how they are enforced,
in all its forms for many purposes. Moreover, homeless and the activities of homeless people on the real and
people are adept at overcoming economic and technological perceived safety of city dwellers, including themselves, are
barriers to obtain access [16,26]. Additionally, through the complex. Here, we do not explicate them; that said, we do
use of digital technology in public and semi-public settings, make two basic claims that help frame our work. First,
tensions among the homeless and other urban dwellers can homeless young people, in the struggle to meet basic needs,
become prominent [18]. The need to connect personal frequently encounter unsafe situations, with civility laws
devices to infrastructure, for example, to recharge batteries, often being implicated. Enforcement against sleeping in
may lead the homeless to come together into postures that public parks, for example, may lead young people to less
are perceived as threatening by some people. We interpret safe sleeping situations. Second, the desire, and
these findings to suggest that the widespread diffusion of increasingly the need, to access digital infrastructure can
digital and mobile technologies into homeless communities also bring young people into unsafe situations. The need for
is changing the nature of homelessness in ways just being electricity to recharge a phone, for example, may lead
discerned, with potential opportunities for equalization young people to trespass at secluded power outlets [24].
across social and economic classes, along with dangers of Value Sensitive Design
further stigmatization and entrenchment. To enable a comprehensive analysis of safety and other
Another area of related work concerns a new class of value tensions at the research site, we drew on Value
mobile applications for improved safety of “homed” Sensitive Design [11,12,20,21], with the key theoretical
teenagers [4]. In short, these applications enable parents to constructs and methods used in this work described here:
monitor their teen’s activities, providing information on Stakeholders. Value Sensitive Design makes a distinction
what they are reading online, where they go, how they between direct and indirect stakeholders. The former
drive, who they are with, and so forth. By the availability of interact directly with the system; the latter while not
such intimate information, parents are called upon to interacting directly are affected by the system’s use [11,12].
balance their desire to keep their children safe with other In the framing of this research, we assumed that homeless
values related to the raising of children including trust, young people are the primary direct stakeholders who will
privacy, autonomy, independence, and maturation. adopt mobile phone technologies for keeping safe. We also
Place, Homelessness, and Safety assumed that other direct stakeholders might emerge as the
As noted above, tensions arising from the use of mobile work progressed. In addition, in prior stakeholder analysis
devices by the homeless are a contemporary example of a [28] we identified three key groups of indirect stakeholders:
long-standing theme. In short, since Elizabethan times, the service providers, police officers, and community members.
“homed” have often viewed the “homeless” with Importantly, we also assumed that under certain
trepidation. Consider, for example: circumstances indirect stakeholder groups might take on the
role of direct stakeholders [4,8,9]. However, in keeping
In the western world ‘Home’ is an ideal as well as a place – a
spatially constructed ideology usually correlated with with our emphasis on homeless young people as the primary
housing…. homelessness also signifies ‘displacement’ – an direct stakeholders, in the work reported here we asked all
existential lack [of identity and morality] that is perhaps even participants to consider how mobile phones might be used
more fundamental than being without shelter” [3, p. 115]. to keep a homeless young person safe.
In other words, in the struggle to meet basic needs, Value Tensions. Prior work [5,20] has identified the
homeless people are often forced to work against, or at least importance of engaging value tensions. In our work with

1708
CHI 2011 • Session: Homeless Users May 7–12, 2011 • Vancouver, BC, Canada

homeless young people we anticipated complex value consequence of these considerations, the data we collected
tensions within an individual (e.g., a homeless young are not always completely comparable across all
person both wanting to be self-sufficient on the street and stakeholder groups. Accordingly, we are careful to make
wanting to be cared for by service providers), with an appropriate qualifications as needed.
individual’s relationship to a stakeholder group (e.g., at
The Research Setting: The U-District, Seattle, WA
times a homeless young person might feel threatened or less
This research was conducted in the University District,
safe around police and at other times the same person might
hereafter the U-District, a neighborhood located adjacent to
call upon police for help in an unsafe situation), and among
the University of Washington. Since the 1960s, community
different stakeholder groups (e.g., when homeless young
members have developed an urban place for welcoming
people’s needs for physical safety, such as using a doorway
“wanderers” with free medical, counseling, and shelter
of a store in order to stay dry, are in tension with
services. Today, an alliance of nine agencies provides a
community members’ feelings of safety in public).
continuum of care for homeless young people [25].
Value Sketches. For people who live in the public and must
Researcher Stance
frequently move to obtain basic needs, perceptions of safety
The research team is comprised of people with backgrounds
are largely situated in place and time of day. To better
in HCI, design, and security. We work from a design stance
understand this way of life, we employed a sketching
that seeks to both understand the situated context as well as
activity, called value sketches, which prompt participants to
to create meaningful change. Because of our prior volunteer
represent place, mobility, and safety. Sketching is a
work at a community technology center [26], we were able
common qualitative method for uncovering knowledge for
to quickly establish a trusted rapport with participants –
physical and conceptual structure [10,13,15,19,23]; value
especially important when asking questions about “safety.”
sketches in particular emphasize participants’ values and
In varied ways our research efforts are connected with our
involve a systematic analysis of the drawn elements.
collaborating organizations, the young people they serve,
Value Scenarios. While value sketches are well suited to and the greater U-District. On reflection, we have also
elicit knowledge situated in place, not all types of found that this work changes us, especially how we
knowledge and feelings can be most readily expressed apprehend the U-District and think of the young people we
through drawings. Stories that emphasize social and value have met in research and later encounter on the street. In
considerations of new technologies, called value scenarios these public encounters, we are trained to acknowledge the
[4,21], can fill this gap. For foregrounding value person but subtly, so as not to reveal our inter-personal
implications and envisioning systemic effects, value familiarity to bystanders.
scenarios were first proposed for use by designers as an
extension to use and problem scenarios [2]. Here, we Participants, Recruitment and Data Collection Contexts
evolve the use of value scenarios by placing their creation The recruitment and interviewing procedures varied
in the hands of the participants to elicit ideas for how somewhat by stakeholder group as follows: (1) Homeless
mobile phones could be used to improve the safety of young people (14 men, 5 women; ages 19–32, M=24).
homeless young people. Service providers at one of the local agencies in the U-
District recruited two same-sexed groups of homeless
METHODS young people with group interviews taking place on
Methodological Considerations separate days at a well known drop-in facility. (2) Service
Given the complexity of homeless young people’s providers (4 women; ages 21–41, M=29; 2, 4, 36, and 72
circumstances introduced above, we took a deliberately months experience). Employees at service agencies who
exploratory approach. We did not want to commit to tightly work with homeless young people were recruited by
focused research instruments; rather, we sought to develop personal contact and interviewed as a group in their place
instruments that would allow us to bound the design space of work. (3) Police officers (1 man, 1 woman; ages 38;
broadly and that would lend themselves to adaptation as the M=38; 4 and 12.5 years experience). University police
work unfolded. For instance, we did not define the value officers who have patrolled the U-District were recruited by
“safety,” instead we provided instruments which we hoped personal contact and interviewed in their place of work.
would afford the opportunity for participants to reveal their And (4) community members (14 men, 4 women; ages 18–
own perceptions and experiences of safety. In addition, we 84, M=52). Other people who attend school, visit, live, or
sought to involve a wide spectrum of stakeholders, work in the U-District were recruited and interviewed
including homeless young people, service providers, police individually at the annual U-District Street Fair.
officers, and community members. Thus, we needed to
Special considerations for working with homeless young
accommodate their unique characteristics by collecting data
people. Homeless youth self-selected to be included in this
in different contexts. Finally, while the empirical work was
research. To protect their identities they provided oral
to provide significant stand-alone data on homeless youth,
assent and participated under the auspices of a collaborating
mobile phones, and safety it was also undertaken with a
service agency. Older participants (into their early 30s) who
view toward informing future co-design activities. In
identify with the community of homeless young people

1709
CHI 2011 • Session: Homeless Users May 7–12, 2011 • Vancouver, BC, Canada

were welcomed to engage in the research activities. developed from the homeless young people data and
Following norms for remuneration at the research site, the applied to the data for other participants. To test the inter-
homeless young people participants received a $25 gift card rater reliability for the value sketches, a second independent
(other participants were not compensated). coder was instructed in the coding manual and re-coded all
Procedures the sketches. Inter-rater reliability was assessed using
Participants in each stakeholder group completed, in this Cohen’s kappa, a measure of the level of agreement
order, a value sketch, a paper and pencil survey, and a value between two coders, with κ = .788. Inter-rater reliability for
scenario; and engaged in a semi-structured discussion the value scenarios followed a similar procedure, with κ =
concerning safety in the U-District. .815. Two commonly referenced benchmarks for
interpreting the values of Cohen’s kappa are Fleiss [7], who
Value Sketches. To gain insight into participants’ rates any value of κ over 0.75 as excellent agreement,
perceptions of safety in the U-District for homeless youth, between 0.40 and 0.75 as intermediate to good, and below
participants completed a value sketches activity, where they 0.40 as poor; and Landis and Koch [17], who rate a κ of
used green and red markers to indicate “safe” and “unsafe” 0.81 to 1.00 as “almost perfect” and between 0.61 and 0.80
areas respectively on a 17×20 inch map of the U-District. as “substantial” agreement.
Of note, community members’ value sketches did not
address their perceptions of safe and unsafe places for RESULTS
homeless young people; thus, we do not report on value Stakeholder-Specific Surveys: Participant Backgrounds
sketches for community members here. To provide a context for understanding the data, we first
establish that all stakeholder groups have some prior
Stakeholder-specific Surveys. To collect information on interaction with homeless young people and that the
demographics, mobile phone use, and related topics homeless young people we worked with had prior mobile
participants completed a 3-page survey, with stakeholder- phone experience. To begin, the service providers reported
specific questions. The surveys included questions extensive ad hoc and regular weekly face-to-face
primarily on cell phone use (e.g., Homeless young people interactions. Complementing their face-to-face exchanges
were asked “Have you ever owned a cell phone?”). were a variety of technologically meditated ones, including
Value Scenarios. To elicit ideas for how a mobile phone email, landline and mobile phone, MySpace or Facebook,
could help homeless young people stay safe, participants and text messaging. The university police officers (16 years
were instructed to write a value scenario, with this prompt: of combined experience) reported they generally interacted
Homeless youth and young adults may face special with homeless young people 1–10 times in a week. These
challenges in keeping safe from harm. Please write a story face-to-face interactions included both ad hoc friendly
about how a cell phone could help to keep a homeless youth social interactions and responding to complaints filed about
or young adult safe. There are no right answers. The story or by homeless young people. In contrast, only one-third of
can be as long or short as you like. It can be about a real the community member participants reported prior
situation or about a fictional situation. interactions with homeless young people, such as being
Importantly, these instructions ask for real or fictional asked for spare change.
scenarios, with the rationale that if participants could mask In terms of prior mobile phone use, 11 of the 19 homeless
some of the facts in their stories they might be more likely young people currently own a phone (length of ownership:
to reveal intimate, subversive, or otherwise risky, though 2 days to 7 years; M=1.5 years) and 5 additional young
plausible, ideas about the use of mobile phones. people have owned a phone in the past. Additionally, more
Discussions. Finally, to provide an opportunity for open- of the young women (100%) than young men (43%)
ended conversation about safety for homeless young currently own a phone (Fisher’s exact test, p = 0.045).
people, participants engaged in a discussion guided by a set
Value Sketches
of 15 questions, which were tailored to the stakeholder
The value sketches provide a window into how participants
groups (e.g., “What kinds of things or circumstances make
viewed the physical space of the U-District in terms of safe
[you / homeless young people] feel [safer / less safe] in the
and unsafe places for homeless youth. The data is anchored
U-District?” and “Imagine a situation in which [you / a
in place, highlighting spots, paths, and regions. Daytime
homeless youth or young adult] might feel less safe. Can
and nighttime maps provide a means to understand the
you think of a way that a cell phone could help?). For the
impact of daylight/daytime-activities vs. dark/nighttime-
community members, the semi-structured discussions (20–
activities on perceptions of safety.
60 min.) were one-on-one and hand-written notes were
taken; for all other stakeholder groups, the semi-structured A total of 50 maps (one daytime and one nighttime map for
discussions (60–90 min.) were audio taped and transcribed. each of the 19 homeless young people, 4 service providers,
and 2 police officers) were analyzed. To analyze these maps
Coding and Reliability. Drawing on prior work [9,12], we
and as shown in Table 1, our coders were instructed to
developed coding manuals for both the value sketches and
begin with the identification of spots (Category I), specific
scenarios (see Tables 1 and 2). Coding manuals were
locations that people can walk to; paths (Category II), along

1710
CHI 2011 • Session: Homeless Users May 7–12, 2011 • Vancouver, BC, Canada

I. SPOT III. REGION


A. Private A. Unlabeled
1. Labeled B. Labeled
2. Unlabeled 1. With safety focus (see Fig. 1d)
B. Public 2. Without safety focus
1. Bar, café, etc.
IV. COMPOSITE
2. Church, etc.
A. Listing
3. Park
B. Nested
4. Service agency
1. Exceptions (see Fig. 1e)
(see Fig. 1a, 1b)
2. Elaborations
5. Municipal building
6. Other place
II. PATH
A. Unlabeled (see Fig. 1c)
B. Labeled
1. With safety focus
2. Without safety focus
Table 1: Value sketches codes. (Uncodeable codes not shown.)
which people can walk; and regions (Category III), within
which people can walk to many spots and along many
paths. These features, each of which can be seen in Figure Figure 1: Value sketch (daytime), showing two spots (a and
1, were expressed through the varied use of graphic marks b), a path (c), a region (d), and a nested exception (e), made
(e.g., spots were expressed by Xs, circles around a building, by a homeless young person (cropped image). Green and
or other shapes, etc.). Overall there were a total of 146 red represent safe and unsafe features, respectively.
spots, 68 paths, and 55 regions marked on the 50 maps.
with which homeless young people perceive safety in the U-
Labels on the maps served multiple functions, at times District during daytime and nighttime.
acting to identify landmarks and at other times to indicate
In terms of location, against a background of substantial
temporal, person-based, or activity-based qualifications of
variation in the features (e.g., many features labeled on the
safety. Some participants made use of pre-existing labels on
maps were unique to a single individual), a large amount of
the map (e.g., University Bookstore); other participants
consistency emerges in how the young people view most
wrote their own annotations, which were often placed in the
types of locations as either safe or unsafe. For example, the
maps’ margins. Such linguistic, along with pictorial and
homeless young people generally consider service agencies
iconographic, marks were associated by proximity,
to be safe places; service agencies were marked a total of
enclosure, and connectivity. For example, a green drawing
32 times and indicated safe in 28 cases (88%). Interestingly,
of trees, labeled
while both young women and men marked service agencies
“Ravenna” (the
as safe spots, the young women marked service agencies as
name of a local
safe spots on their maps more frequently than the men (Z =
park) indicates a
-3.52, p = .0004, Mann-Whitney U test). Similarly,
safe location whereas a red foamy beer mug,
churches were indicated safe in 10 of the 11 cases (91%)
labeled “drunken danger” placed near a bar
where they were marked. In terms of unsafe locations, bars
indicates an unsafe location.
were marked a total of 8 times, always indicated unsafe
Participants often employed graphic marks in combination, (100%). The youth also seemed wary of college students,
at times placing spots within paths, spots and paths within with the university campus indicated unsafe on 6 of the 7
regions, and regions within regions. These composite visual maps (86%) where it was marked and “frat row” indicated
structures (Category IV), along with linguistic marks, unsafe on all 4 maps (100%) where it was marked.
provide a good deal of expressive power for indicating In terms of time of day, we expected homeless young
information about place-based safety. For instance, “COP people to view the U-District as safer during the day than
SHOP,” depicted as a red rectangle, is placed on a major during the night. Specifically, we expected young people to
thoroughfare, depicted by a thick green line (see Figure 1e). indicate fewer unsafe locations on their daytime than
Given our emphasis on homeless young people as the direct nighttime sketches. There is some suggestion that this is the
stakeholders in this research, we now turn to two aspects of case, with 29 of the 100 (29%) features marked on the
their sketches: location, the consistency with which daytime maps indicated unsafe as compared to 39 of the 93
homeless young people view specific places in their (42%) nighttime features indicated as unsafe, but the
environment as safe or unsafe; and time, the overall patterns difference is not statistically significant. In addition, three
young people marked the entire U-District as safe during

1711
CHI 2011 • Session: Homeless Users May 7–12, 2011 • Vancouver, BC, Canada

I. SITUATION III. MOBILE PHONE TECHNOLOGY H1. If you need help you can call for it. If your car runs
A. Reactive A. Functionality out of gas you can call your friends to bring you some
1. Hostile 1. Traditional calling gas. And a whole bunch of other sh*t too. [Male]
2. Accident 2. Customization H2. Some friends of mine had gone to get more beer.
3. Authority 3. Sensors My phone rang, my homies called to tell me the boys
4. Environmental 4. Recording were heading my way. I had enough time to put my
B. Preventative 5. Technical network weed away before they arrived. [Male]
C. Secondary effect 6. Information seeking H3. I would use devices in my cell phone to record law
7. Social networking enforcement, when they choose to harass me. [Male]
II. PURPOSE
B. Form
A. Warnings H4. I don’t think cell phones keep people safe because if
1. Material
B. Calling for help you call the cops for seeing a crime you might get beat
2. Shape up later for snitching. [Male]
C. Maintain relationships
3. Visibility
D. Documenting H5. I feel when hitching rides, with a cell phone you can
C. Monetary costs
E. Information gathering be kept safe. If you’re walking down the road with your
F. Leveraging social expectations IV. LOCUS OF WELFARE thumb out and a cell phone to your ear a “weirdo” is less
G. Preventative A. Self likely to pick you up. [Female]
B. Other
H6. Once upon a time there was three little pigs, one
lived in a house, one lived on the street, and the last one
Table 2: Value scenario codes. Italic codes identified
lived in a squat. One day a big bad wolf was looking for
conceptually but not used. (Uncodeable codes not shown.)
a squatter, the big bad wolf was out to get all the little
the daytime (possibly with a few exceptions), but did not do pigs. The first little pig called the second pig, and he
found the third pig through word of mouth. Thank
so on their nighttime maps, perhaps indicating that the areas
cellphone. [Male]
where they feel safe are more localized at night.
H7. Usually if you have a phone people know the police
Value Scenarios are on the way. If someone is using an expensive
The value scenarios take off from where the value sketches phone, it might escalate the situation. [Male]
end. Whereas, the value sketches provide a good deal of C1. Perhaps they might be a bit safer w/a cell phone if
insight into participants' nuanced conceptions of the U- someone were to take advantage of them, they could
District as a place with safe and less safe physical spaces call police. The question is would the police come to
for homeless young people; the sketches tell us less about his/her aid? [Female]
the social relationships within those spaces, subtleties tied P1. Being homeless is a challenging experience.
to the situation and contexts of interactions, the nature of Without a job and access to food, shelter, & other basic
people’s purposes, and the specific role of mobile phone necessities, living on the streets is emotionally &
technology including its functionality, form, and material physically draining. Having a cell phone would at least
allow me to be able to call other homeless persons that I
costs. Participants largely freeform value scenarios brought
have networked with to work on attaining these difficult
forth all of these and other considerations. to get necessities. Knowing that I could call a shelter to
Table 3 shows 10 illustrative scenarios, selected mainly for get a bed, or 911 for an emergency would make me at
their coverage of the coding manual, from the 43 collected least feel safer and closer to services that I need while
living on the streets. Others can check on me as well to
(19 homeless young people, 4 service providers, 2 police
make sure that I am safe throughout the day. [Male]
officers, and 18 community members; scenario length: 5-
150 words, M=44 words). As shown in Table 2 the analysis S1. A man was harassing a youth in her sleeping spot.
She was able to snap a picture of the man and forward it
yielded four distinct dimensions: situation, purpose, mobile
to service providers & police. With this photo the man
phone technology, and locus of welfare. Each scenario was was picked up and charge [sic]. Hopefully he will not be
coded for each of the four dimensions. Complex scenarios sexually harassing any more homeless young women.
received multiple codes within each dimension to reflect the [Female]
depth and nuance of the scenario. Given our exploratory
goals, we sought to uncover statistical differences in the Table 3: Value scenarios, labeled by stakeholder group:
Homeless Young People (H1-H7), Community Members (C1),
coding of the scenarios based on stakeholder groups or
University Police (P1), and Service Providers (S1).
gender. However, perhaps due in part to the small sample
warn others of an impending event (II.A), to call for help
sizes, no statistically significant differences were found.
(II.B), to document an event (II.D), or to actively leverage
Situation refers to elements in the environment that suggest social expectations (II.F). Mobile Phone Technology refers
or explain why the mobile phone is used; for example, in to the technology used in the scenario; for example to the
reaction to a hostile event (Code I.A.1) or accident (I.A.2), phone’s functionality in making calls (III.A.1) or recoding
to prevent an anticipated event from occurring (I.B), or due audio or video (III.A.4), to the phone’s form such as its
to secondary effects that followed from the presence of the shape (III.B.2), or the costs entailed in owning and using a
mobile phone (I.C). Purpose refers to the protagonist’s mobile phone (III.C). Lastly Locus of Welfare refers to
aims for taking action with a mobile phone; for example, to

1712
CHI 2011 • Session: Homeless Users May 7–12, 2011 • Vancouver, BC, Canada

whom the protagonist directs a concern for safety, to him or intelligence systems with “internet profiles and database
herself (IV.A) or toward others (perhaps including the access” to plan for improved safety (III.A.5), and imagined
protagonist) (IV.B). Overall, there were a total of 50 the materiality of the phone being used to “save your life by
situation, 49 purpose, 51 mobile phone technology, and 43 jumping in front of a bullet for you” (III.B.1).
locus of welfare codes for the 43 scenarios. Scenarios surfaced not only direct impacts from the mobile
To provide a flavor for how these dimensions and the phone but secondary ones as well that typically resulted
categories within them characterize the scenarios, consider from the form or monetary cost of the phone. One homeless
some of the exemplar scenarios presented in Table 3. For young woman wrote in H5 that when hitchhiking, visibly
example, scenario H1 was coded as reactive/accident holding “a cell phone to your ear a ‘weirdo’ is less likely to
(I.A.2) as the protagonist faces a situation in which the pick you up.” Here, the perceived benefit to safety comes
mobile phone is used in direct response to “running out of from others who might cause harm “seeing” that the
gas;” calling for help (II.B), indicating the protagonist’s homeless young person could contact someone via phone;
purpose in using the mobile phone to “call your friends to not that a particular phone call has been made.
bring you some gas;” traditional calling (III.A.1), to Finally, not all uses of mobile phones were perceived as
indicate the technical aspect of the mobile phone that is improving safety (H4, H7). For example, in H4 a homeless
employed to make a “call;” and self (IV.A), to indicate that young man suggests that in reporting a crime to the police
the scenario’s protagonist is concerned with his own safety “you might get beat up later for snitching.” In H7, another
as in “if you [the protagonist] need help.” young man calls attention to the risks to safety from
While allowing for rich diversity in the details and as possessing relatively expensive technology.
exemplified by H1 above (see also C1, P1), one common
DISCUSSION
narrative structure emerged from the data and accounted for On the Meaning of “Safety”
roughly half (49%) of the scenarios. Perhaps not Recall that we did not define “safety;” rather we presented
surprisingly and as reflected in the coding manual, this participants with open-ended yet structured activities to
narrative structure describes reactive situations (I.A) in elicit their perceptions and experiences. Participants told us
which the protagonist seeks help (II.B) by using a mobile through sketches and stories the myriad of places,
phone to make a traditional phone call (III.A.1) for his or situations, and people who come together to make homeless
her own benefit (IV.A). Eight of the scenarios were coded young people more or less safe. From these materials, we
with exactly this group of four codes, and an additional 13 have seen that safety is fundamentally situational. Homeless
scenarios were coded as minor variations that included at young people face both ordinary and extraordinary risks –
least three of these four codes. That said within this running out gas (scenario H1) or getting lost in the
common structure, scenarios portrayed diverse situations mountains, along with chronic exposure to the possibility of
and conveyed rich details from “staying on the street” harassment from police and community members (H3, S1),
where homeless young people are “harassed,” “beat up,” or together with the possibility of isolated, severe violence
“raped,” to protecting “my squat” from the police, to (H4). The police are perceived to be a threat, but so too are
getting “lost in the mountains,” or to “running out of gas.” community members, service providers, and other homeless
Other less frequently occurring patterns were also young people. At the same time, and herein lies a major
identified. For example, the codes documenting (for conundrum, the data also show that youth look to these
purpose) and recording (for functionality) were both rarely same stakeholders and places, at least at times, for safety.
used, but tended to occur together in cases where they were Within this context, the coding manuals provide some
used (p = .0008, Fisher’s exact test). Scenarios H3 and S1 specific dimensions for characterizing the uses of mobile
exemplify such a case, in one instance recording harassment phones for keeping safe.
from law enforcement (H3) and in the other harassment
On Distancing Mechanisms: Sensitivities around Safety
from a man on the street (S1). Only 5 scenarios were
For many people, safety is a highly sensitive topic.
assigned the code recording, and 3 of the 5 were also
Particular details about keeping safe might put a youth or
assigned the code documenting; these were also the only 3
his or her friends at risk, be embarrassing, or even traumatic
scenarios that were assigned the code documenting. That is,
to recall. We might expect, therefore, that some people
scenarios with the purpose of documenting always
might try to create distance between themselves and their
mentioned a recording function, showing how the coding
views on safety, perhaps by obscuring some facts, using
manual represents purposes and functionality.
ambiguity, or stepping away from specificity. In a telling
Working within the structured but open-ended value example of this sensitivity, which arose during the group
scenario format, participants were generative and discussion, one young person said: “Letting people know
imaginative in linking technological opportunities to where I feel safe, makes me feel less safe.”
problematic situations. For example, in their value
Accordingly, we intentionally employed methods that
scenarios homeless young people proposed “map apps to
provided the means for participants to control both the
find safe places” (III.A.6), envisioned “self aware” artificial

1713
CHI 2011 • Session: Homeless Users May 7–12, 2011 • Vancouver, BC, Canada

precision and ambiguity of the information provided. For “homie” or the “police?” Young people must work within
example, in the value sketches participants could provide this uncertainty, leading to the importance of secondary
more or less specific information about location (e.g., effects, leveraging social expectations, and form.
labeled or unlabeled, spot vs. region). Moreover, in the Homeless young people have complex relationships with
value scenarios participants were not asked to differentiate the service providers. While young people may look to the
between real and fictitious but plausible situations. service providers for help, they also may react negatively to
For both methods, indeed, it appears that some homeless particular responses and rules, as might a “homed”
young people did in fact create distance between their own adolescent to his or her parent. We saw, for example, that
experiences and what they were comfortable sharing with all the young women marked service agencies as safe on
us. In the value sketches, three participants marked places, their value sketches; however, in discussion they also
otherwise mundane, as safe or unsafe without indicating expressed frustration with particular service providers’
why (Category I.A.2). These were puzzling. Perhaps, we actions. In a similar sentiment, a young man said: “The
conjectured, the marks indicated locations of one-time service providers sometimes don’t make me feel safe… a
altercations or places related to safety that should otherwise lot of people think they are all friendly… but it turns out
remain unelaborated. If so, with more focused procedures, they are harsh.”
it might be possible to elicit greater detail about these spots. The relationship between homeless young people and
However, paradoxically, by seeking to elicit greater detail, police officers is similarly ambivalent. While the police
young people might be unwilling to say anything at all. We were often seen to pose a threat (H2, H3, H6), they were
conclude, therefore, that some marks on the value sketches also seen as potentially beneficial (H7, C1, S1). While
might be a form of “distancing,” that is, conveying young people’s sketches do not contain a single positive
something important while keeping oneself out of it. statement about the police (but a fair number of negative
Further examples of distancing might be present in the statements), young people in discussion and in the scenarios
value scenarios, where narrative forms can convey essential did indicate that they would call upon the police and 911 in
information but without indicating a personal connection to some emergences. At the same time, while police officers
the events, protagonists, or conflicts. Scenario H6, a “fairy conveyed a certain dislike toward homeless young people
tale” about three little pigs and a big bad wolf, was such an (“It’s like they own the parking lot, the donut store, the
example. Interestingly, the basic content and structure of [sandwich shop],”) they also appreciated the difficulty of
this scenario – namely, that cell phones can be used to warn surviving on the streets (P1).
of an impending police presence – can be conveyed through Finally, homeless young people saw different groups of
different forms of writing. Scenario H2, in “realistic” style, community members as safety risks. One group, called
illustrates the contrast quite well. “angry Christians,” was described as intolerant and violent.
In a different vein, participants might create distance Concerning fraternity members a young person said: “You
between themselves and their experiences by the use of ever woke up to them [frat kids] rummaging through your
language, by, for example, placing events in the passive stuff and then kicking you in the face when you said ‘Get
voice or in hypothetical terms. While we have not out of my stuff.’” Recall too that “frat row” and regions of
completed a close analysis of how language might be used the university campus were indicated to be unsafe in the
to “create distance” we speculate that by the use of such value sketches. Community members used words such as
conditional words as “if,” “would,” and “could” “squatters,” “street kids,” and “Ave rats” to name the
participants might remove themselves from particular people who were perceived to make the U-District unsafe.
situations but convey their essential content and structure. Like the police officers, however, the scenarios showed
This linguistic analysis remains for future work. community members to be empathetic (C1), with some
Value Tensions scenarios envisioning communication networks specifically
For considering “relationship,” instead of dichotomies or designed to keep homeless young people safe.
conflicting points of view, we prefer “value tensions.” Design Opportunities
Value tensions capture the mutuality of relationship; they When discussing the limitations of mobile phones and when
are not necessarily problematic, nor do they always cross a justifying new possibilities, the stakeholder groups brought
contentious space. Instead, tensions can keep two points of forward a variety of functional criteria, operating
view together in balance or enable two conflicting views to constraints, and general desiderata for evaluating features.
coexist. This analytical orientation, in short, allows us to The general desiderata included: (1) low cost; (2) outdoor
present and reason about relationship with nuance. durability for repeated drops, wet, and cold; (3) flexible
While homeless young people might help each other keep powering options, decoupled from infrastructure (e.g., solar
safe (H1, H2, H6, P1), by using a mobile phone to call for recharging); (4) 24/7 reliability for emergency use; (5)
help, they may also be seen as “snitching” (H4), a violation independent and separate communication channels,
of the rules of the street. The use of a phone in many protected from surveillance; and (6) the possibility for
situations is ambiguous – is that young person calling a recovery when lost by inattention or stolen. These grounds

1714
CHI 2011 • Session: Homeless Users May 7–12, 2011 • Vancouver, BC, Canada

for judging features, quite obviously, come from the Technical features: Provide access points to electricity and
circumstances of homelessness and provide a backdrop for wireless capability at sites not considered unsafe.
the following opportunities for design: Moreover, enable access points to be moved to different
Support for shifting trust relationships among homeless locations so that no one location in the U-District becomes
young people and enabling young people’s desire to help known as “that place for the homeless.” One possible
others (H1, H2, H6, P1). While young people face an solution, or at least design direction, is large-sized tricycles
environment of shifting alliances, they also come to trust [27] that act as mobile power and Internet connectivity
and rely on each other to respond to calls for help. When sources, placed at suitable locations for all urban dwellers.
young people are in trouble, they could benefit from ways Such infrastructure might mitigate the risk that homeless
to reach out to trusted peers or groups, albeit the specific young people experience additional forms of stigma by the
individuals might change quite frequently. Technical need for infrastructure.
features: Provide capability to represent alliances, to CONCLUSIONS AND CONTRIBUTIONS
diplomatically bring people into and out of them, to phone Implications for Co-Design. The coding manuals for the
specific contacts, to broadcast calls for help or warnings, value sketches and the value scenarios shape a design space
and to communicate without the possibility of surveillance. for further investigation of design opportunities. We intend,
Support a homeless young person’s need to document for example, to use the scenarios as prompts for considering
abuse (H3, S1). Homeless young people described specific design solutions and to use the coding manuals to
occurrences of harassment and violence, from police situate particular solutions. In addition, we expect that value
officers and other community members, as well as other sketches, together with particular prototypes, will enable
homeless people. Documenting such events can empower homeless young people and other co-designers to enact and
young people to seek redress and, over time, may deter envision the consequences of particular designs in action.
future harassment. Technical features: Provide capability to Contributions. In addition to the design opportunities, we:
document events in real-time (e.g., audio and video (1) Documented rich knowledge about the interaction
recording) as they unfold, and to place such documentation among place, mobile phone technology, and safety for a
into the hands of someone who or an institution that can historically underrepresented population, homeless young
facilitate an appropriate response. people; (2) Demonstrated the use of innovative open-ended
Support a surreptitious call for help (H4, H7). Being seen yet structured methods to elicit views on a sensitive topic
or “perceived” as making contact for help or documenting a (personal safety) in a non-threatening and dignified manner,
situation as it unfolds can put a homeless young person at leaving what information to reveal and how under the
risk, as telling of street activities to outsiders can be seen as control of the participant; (3) Constructed coding manuals
a form of weakness or “snitching.” Technical features: that can be used and extended by others to analyze the
Provide capability to call and/or document a situation relationships among place, mobile phone technology, and
without being visible (e.g., eyes-free “panic” button), to safety; and (4) Extended a method (i.e., value scenarios
represent emergency networks or trusted others, and to generated by participants) in Value Sensitive Design,
diplomatically bring people into and out of those networks. thereby contributing to that growing literature.
Support for homeless young people as urban dwellers with Final Words. Surviving on the street requires that young
specific information needs. The value sketches and people develop expertise for managing unsafe situations;
scenarios capture something of the distinct perspective that mobile phones are surely implicated because of their
homeless young people have for their physical environment pervasiveness and because they are worn on bodies. By
(e.g., churches as places of safety, areas of parks safe at investigating this setting through the systematic and
different times, etc.). Technical features: With location- synthetic analysis of place-based representations (value
based services, provide the capability to represent specific sketches) along with narratives of possible uses of mobile
kinds of information that homeless people and those living phones (value scenarios) we have been able to more fully
in poverty might seek alongside other information about a explore the design space of mobile phones and safety. In
city. Addressing this goal will foreground tensions between some ways, elements of this setting are at the same time
homeless communities and the other stakeholders [25]. both extraordinary and ordinary [26]. Thus, much of what
we have uncovered in this extraordinary setting, along with
Support for safe, non-stigmatized access to infrastructure. the methodological approach and methods employed, is
The value sketches alerted us to perceptions of safe and likely to be applicable to other settings and circumstances
unsafe areas in the U-District. Recall, for example, that the where people look to their mobile phones for safety.
university grounds, areas of student housing, and the
neighborhood service center (i.e., “cop shop”) were ACKNOWLEDGMENTS
considered generally unsafe by young people. At the same Special thanks to all participants for their willingness to
time, the discussions based on the value scenarios pointed consider safety and mobile phones. Thanks to Sanjana
to the need for access to electricity to recharge batteries. Prasain for help in coding the value sketches. This research
was supported in part by NSF Award CNS-0905384.

1715
CHI 2011 • Session: Homeless Users May 7–12, 2011 • Vancouver, BC, Canada

REFERENCES 15. Hendry, D. G., and Efthimiadis, E. N. Conceptual


1. Beckett, K., and Herbert, S. Banished: The New Social models for search engines. In Web Search:
Control in Urban America. Oxford University Press, Interdisciplinary Perspectives, A. Spink and M. Zimmer
New York, 2010. (Eds.). Springer, Berlin, 2008, 277-308.
2. Carroll, J. M. Five reasons for scenario-based design. In 16. Karabanow, J., and Naylor, T.D. “Being hooked up.” In
Proc. HICSS 1999, IEEE (1999), 3051-3061. Digital Diversity, E. D. Looker, and T. D. Naylor
3. Cresswell, T. Place: A Short Introduction. Blackwell (Eds.). Wilfred Laurier University Press, Waterloo, ON,
Publishing, Oxford, 2004. 2010, 161-178.
4. Czeskis, A., Dermendjieva, I., Yapit, H., Borning, A., 17. Landis, J., and Koch, G. The measurement of observer
Friedman, B., Gill, B. T., and Kohno, T. Parenting from agreement for categorical data. Biometrics 33 (1977),
the pocket: Value tensions and technical directions for 159-174.
secure and private parent-teen mobile safety. In Proc. 18. Le Dantec, C. A., Christensen, J. E., Bailey, M., Farrell,
SOUPS 2010, ACM Press (2010), Paper 15. R. G., Ellis, J. B., Danis, C. M., Kellogg, W. A., and
5. Denning, T., Borning, A., Friedman, B., Gill, B., Edwards, W. K. A tale of two publics. In Proc. DIS
Kohno, T., and Maisel, W. Patients, pacemakers, and 2010, ACM Press (2010), 11-20.
implantable defibrillators: Human values and security 19. Lynch, K. The Image of the City. The MIT Press,
for wireless implantable medical devices. In Proc. CHI Cambridge, MA, 1960.
2010, ACM Press (2010), 917-926. 20. Miller, J. K., Friedman, B., Jancke, G., and Gill, B.
6. England, M. Stay out of drug areas: Drugs, othering and Value tensions in design: The value sensitive design,
regulation of public space in Seattle, Washington. Space development, and appropriation of a corporation's
and Polity 12, 2 (2008), 197-213. groupware system. In Proc. GROUP 2007, ACM Press
7. Fleiss, J. L., Levin, B., and Paik, M. C. Statistical (2007), 281-290.
Methods for Rates and Proportions, 3rd edition. John 21. Nathan, L. P., Klasnja, P. V., and Friedman, B. Value
Wiley & Sons, New York, 2003. scenarios: A technique for envisioning systemic effects
8. Friedman, B., Borning, A., Davis, J. L., Gill, B. T., of new technologies. Ext. Abstracts CHI 2007, ACM
Kahn, Jr., P., Kriplean, T., and Lin, P. Laying the Press (2007), 2585-2590.
foundations for public participation and value advocacy. 22. Nickelsville (2010). Declaration of a state of emergency
In Proc. dg.o 2008, Digital Government Society of in 2010 by and for homeless people in Seattle & King
North America (2008), 305-314. County. http://www.nickelsvilleseattle.org/
9. Friedman, B., Höök, K., Gill, B., Eidmar, L., 23. Sutton, S.E., and Kemp, S. P. Integrating social science
Sallmander Prien, C., and Severson, R. L. Personlig and design inquiry through interdisciplinary design
integritet: A comparative study of perceptions of privacy charrettes. Am. J. of Community Psychology 38 (2006),
in public places in Sweden and the United States. In 125-139.
Proc. NordiCHI 2008, ACM Press (2008), 142-151. 24. Woelfer, J. P., and Hendry, D. G. Homeless young
10. Friedman, B., Hurley, D., Howe, D. C., Felten, E., and people and living with personal digital artifacts. To
Nissenbaum, H. Users' conceptions of web security: A appear in Proc. CHI 2011.
comparative study. Ext. Abstracts CHI 2002, ACM 25. Woelfer, J. P., and Hendry, D. G. Designing ubiquitous
Press (2002), 746-747. information systems for a community of homeless young
11. Friedman B., Kahn P. H., Jr., and Borning, A. Value people. Personal and Ubiquitous Computing, 2010, doi:
sensitive design and information systems. In Human- 10.1007/s00779-010-0341-5.
computer Interaction in Management Information 26. Woelfer, J. P., and Hendry, D. G. Homeless young
Systems: Foundations, M.E. Sharpe, Armonk, NY, people’s experiences with information systems: Life and
2006, 348-372. work in a community technology center. In Proc. CHI
12. Friedman, B., Kahn, P. H., Jr., Hagman, J., Severson, R. 2010, ACM Press (2010), 1291-1300.
L., and Gill, B. The watcher and the watched: Social 27. Woelfer, J. P., and Hendry, D. G. Stabilizing homeless
judgments about privacy in a public place. Human- young people with information and place. JASIST 60, 11
Computer Interaction 21 (2006), 235-272. (2009), 2300-2312.
13. Gaver, W. W., Dunne, T., and Pacenti, E. Design: 28. Woelfer J. P., Yeung, M. W-M., Erdmann, C. G., and
Cultural probes. interactions 6, 1 (1999), 21-29. Hendry, D. G. Value considerations in an information
14. Gibson, T. A. Securing the Spectacular City. Lexington ecology. In Proc. AM 2008, ASIST (2008), 248-256.
Books, Lanham, MA, 2004.

1716

You might also like