ArticlePDF Available

Creating Stealth Game Interventions for Attitude and Behavior Change: An "Embedded Design" Model


Abstract and Figures

Persuasive games tackling serious issues in a literal, explicit fashion are far less likely to succeed in changing attitudes or behaviors than are games that take the more “stealthy” approach of embedding persuasive messages within a game’s content or context. The “Embedded Design” model, developed by the design and research team at Tiltfactor Laboratory at Dartmouth College, offers novel, evidence-based strategies for including persuasive content in a game in ways that circumvent players’ psychological defenses, triggering a more receptive mindset for internalizing a game’s intended message, and do so without sacrificing players’ enjoyment or the game’s inherent replayability. Such techniques promise to revolutionize the repertoire of techniques that game developers should consider in broaching and presenting serious topics in games. Three original “Embedded Design” strategies are presented here: (1) Intermixing: balancing “on-message” and “off-message” content to render the former less overt or threatening; (2) Obfuscating: using framing devices or genres that divert expectations or focus away from the game’s persuasive intent; and (3) Distancing: employing fiction and metaphor to increase the psychological gap between players’ identities and beliefs, and the game’s characters and persuasive content.
Content may be subject to copyright.
Creating Stealth Game Interventions for
Attitude and Behavior Change
An “Embedded Design” Model
Geoff Kaufman, Mary Flanagan, & Max Seidman
Transactions of the Digital Games Research Association
2016, Vol. 2, No. 3, pp. 173-193
ISSN 2328-9422
TEXT: Licensed under Creative Commons Attribution (CC BY-NC- ND
2.5) nd/2.5/
IMAGES: All images appearing in this work are property of the respective
copyright owners, and are not released into the Creative Commons. The
respective owners reserve all rights.
Persuasive games tackling serious issues in a literal, explicit fashion are
far less likely to succeed in changing attitudes or behaviors than are
games that take the more “stealthy” approach of embedding persuasive
messages within a game’s content or context. The “Embedded Design”
model, developed by the design and research team at Tiltfactor Labo-
ratory at Dartmouth College, offers novel, evidence-based strategies for
including persuasive content in a game in ways that circumvent players’
psychological defenses, triggering a more receptive mindset for internal-
izing a game’s intended message, and do so without sacrificing players’
enjoyment or the game’s inherent replayability. Such techniques promise
to revolutionize the repertoire of techniques that game developers should
consider in broaching and presenting serious topics in games. Three orig-
inal “Embedded Design” strategies are presented here: (1) Intermixing:
balancing “on-message” and “off-message” content to render the former
less overt or threatening; (2) Obfuscating: using framing devices or gen-
res that divert expectations or focus away from the game’s persuasive
intent; and (3) Distancing: employing fiction and metaphor to increase
the psychological gap between players’ identities and beliefs, and the
game’s characters and persuasive content.
Persuasive games, attitude change, behavior change, embedding, game
The past several decades have seen the emergence of a plethora of per-
suasive games that aim to increase players’ awareness of critical and
timely social issues – and to change players’ attitudes and behaviors –
through gameplay (Bogost 2007). Running the gamut from games tar-
geting cognitive biases that reduce the accuracy of judgment and deci-
sion making (e.g., the SIRIUS initiative of the Intelligence Advanced
Research Projects Activity program: Dunbar et al. 2013) to those
intended to encourage behaviors that benefit society (such as recycling
in the case of the mobile game Gaea: Centieiro, Romão, and Dias 2011)
or the self (e.g., the reduction of substance abuse and HIV risk, which is
the focus of the “Play2Prevent” program: Fiellin et al. 2014), this subset
of “serious games” is united by their intention to transform mindsets and
actions through the messages they model.
Games themselves are powerful means of enculturation (Flanagan 2009).
A vast majority of serious games, however, share a common design phi-
losophy: by and large, they present characters, scenarios, situations, and
solutions in a direct, matter-of-fact fashion under the ostensibly logi-
cal (and well-intentioned) assumption that doing so will automatically
encourage and enable players to internalize and transfer the game’s mod-
eled beliefs and behaviors to real-life contexts.
174 ToDiGRA
In light of what is known in contemporary psychology, this approach, we
argue, is ill-advised at best and potentially harmful at worst, particularly
when dealing with persuasive content that is uncomfortable, psycholog-
ically threatening, or counter-attitudinal. A vast body of social psycho-
logical theory and research on persuasion and attitude change has long
demonstrated that it is a basic human tendency to resist persuasive com-
munications that are perceived as too forceful or forthright in their inten-
tions. For one, being aware that some external agent is aiming to change
one’s attitudes or behaviors triggers psychological reactance: an aversive
state of arousal that arises whenever one perceives that his/her freedom
to do or think freely is being threatened (Brehm 1966). The aversive
state of reactance raises individuals’ psychological defenses, rendering
them less receptive (and, indeed, more resistant) to a persuasive mes-
sage. What is even more surprising is that psychological reactance will
occur even if a person’s own beliefs align with the content of the message
(e.g., Worchel and Brehm 1970).
A second psychological barrier that comes into play in situations of per-
suasion and play, especially when dealing with attitudes and behaviors
of a particularly sensitive nature (such as the hot-button issues of stereo-
types and prejudice), is the bias blind spot: the acknowledgment that
biases exist but the denial or minimization of one’s own susceptibility to
those biases (e.g., Pronin, Lin, and Ross 2002).
The potentially aversive and defensive reactions triggered by explicit
persuasive attempts limit not only the potential efficacy of game-based
interventions, but also players’ enjoyment of them, for the perception
of a persuasive agenda is inherently antithetical to players’ immersion
within a game world (and, indeed, antithetical to the notion of play itself:
see de la Hera Conde-Pumpido 2013; Huizinga 1938). In other words,
most persuasive games may fail to engage players, let alone immerse
them in a transformative experience, due to normal psychological human
reactions to overtly “message-driven” interventions.
For this reason, we propose that persuasive games would greatly benefit
Creating Stealth Game Interventions for Attitude and Behavior Change 175
from using a subtler, stealthier approach to presenting their focal mes-
sages or themes. This line of thinking is directly informed by our team’s
research at the Tiltfactor Lab, which has, for over a decade, sought to
tease out effective means for fostering social engagement and enacting
persuasive, prosocial interventions in game design. Under the direction
of Dr. Mary Flanagan, the laboratory team has been building a body
of evidence showing how games can significantly shape perceptions
and change ways of thinking. The lab started by creating more “state-
ment” style games that explored issues as matters for debate or con-
versation; these games were assumed to impact or inform the players
in the same fashion as documentary films and other similar art forms.
Through time, the team has shifted to focus more on an evidence-based
approach to design, using formal experimental methods and a psycholog-
ically grounded approach to demonstrate our games’ impact on players.
As this evolution has occurred, the team has developed novel strate-
gies, including those discussed in this paper, to address controversial
topics, such as public health attitudes and social and cognitive biases,
in a more nuanced, less direct fashion. In this paper, we propose our
novel model of “Embedded Design” that offers key strategies for tack-
ling social issues and including persuasive content in a game in ways
that circumvent players’ psychological defenses, trigger a more recep-
tive mindset for internalizing the game’s intended message, and do so
without sacrificing players’ enjoyment or the game’s replayability.
The key premise of the Tiltfactor Embedded Design model is that the
persuasive impact of game-based interventions is greatly enhanced when
interweaving a focal message within the game’s content, mechanics, or
context of play, rather than making the message or the game’s persuasive
aims the focal point. Through our team’s longstanding work in the design
and study of games intended to shift attitudes and behaviors, we have
uncovered a number of distinct embedding strategies that have proven
effective at increasing our games’ persuasive impact (see Figure 1). This
176 ToDiGRA
work is informed in part by the Values in Design and the Values at Play
methodology, which offers many avenues through which values might
emerge in any given game experience (see Flanagan and Nissenbaum
In this paper, we will focus on three distinct strategies, each representing
a unique manifestation of Embedded Design, that have emerged thus far
in our work: (1) Intermixing: balancing “on-message” and “off-message”
content to render the former less overt or threatening, and more palat-
able and approachable; (2) Obfuscating: using framing devices or genres
that divert expectations or focus away from the game’s persuasive intent;
and (3) Distancing: employing techniques, such as the use of fictional or
metaphorical representations of key issues or themes in order to increase
the psychological gap between players’ identities and beliefs, and the
game’s characters and persuasive content. In the sections that follow, we
expound upon these three strategies and provide concrete examples of
our team’s game designs to illustrate their implementation and cite the
results of empirical investigations that support their efficacy.
Figure 1: The “Embedded Design” Model.
Embedding through “Intermixing”
One means of embedding persuasive content within a game’s design
is a strategy we have come to refer to as “intermixing”: balancing or
interweaving on-topic content with playful but persuasively off-topic (or
off-focus) content that either distracts from the intended message of the
game or helps ease players into the game’s message or aims. This strat-
egy, when implemented effectively, reduces the likelihood of players
experiencing the game as a top-down attitude or behavior change inter-
vention, and offsets the serious (or potentially uncomfortable) tone of the
Creating Stealth Game Interventions for Attitude and Behavior Change 177
“on-message” components of the game with content that has compara-
tively more levity or humor.
Our team has implemented and tested the intermixing strategy in several
game designs. To cite one illustrative example, Awkward Moment (2012)
is a party game for pre- and early-adolescent players that aims to reduce
social biases, including gender stereotypes in science, technology, engi-
neering, and math (STEM) domains (see Figure 2). In Awkward
Moment, players begin with a hand of five “Reaction Cards”; these cards
describe potential responses to the game’s “awkward moments,” includ-
ing actions (e.g., “Scream your head off,” “Write a blog post about it,”
“Talk it out”), exclamations (e.g., “Rats!” “OMG,” “No way!”), and
frames of mind (e.g., “Get serious,” “Relax,” “Channel your inner war-
rior”). During each round, one player serves as the “Decider” and draws
a “Moment Card” that poses a hypothetical situation (e.g., “Somebody
hacks your Facebook account and changes your status to ‘Girls are stu-
pid.’”), to which the other players respond by submitting a Reaction
Card face-down. The Decider then reads each of the submitted cards
and selects a winner for the round. The game aims to stimulate thought
and discussion about responses to social and academic dilemmas, par-
ticularly situations that involve bias against girls and women in STEM.
A subset of the cards in the Moment deck presents situations in which
a female is a target of stereotypes. In some situations, players imag-
ine being a target themselves. The game’s deck of Moment cards con-
tain examples depicting both on-topic scenarios related to gender bias in
STEM (see Figure 2 for an example), as well as off-topic scenarios pre-
senting awkward situations that do not directly pertain to social biases
(e.g., “You sit on ketchup” or “There’s a secret ‘Ugly Poll’ at school, and
you find out you were Number 3 on the list”).
178 ToDiGRA
Figure 2: Sample Reaction Cards (green) and sample Moment Card (blue) from
Awkward Moment.
A key question that guided the iterative design of Awkward Moment was
the ideal ratio of on-topic to off-topic Moments in the game. In line with
the intermixing strategy, our empirical research showed that present-
ing a lower ratio of bias-themed to non-bias-themed Moments proved
much more effective in shifting players’ attitudes and perceptions. One
of our controlled experiments (Kaufman and Flanagan 2015) revealed
that youth participants who were randomly assigned to play an “inter-
mixed” version of the game (with approximately 45% of the Moment
cards depicting gender bias in STEM) exhibited statistically significant
higher post-game levels of perspective-taking, compared to participants
assigned to play an “overloaded” version of the game (with 75% of
the Moment Cards pertaining to bias). In another experiment, an “inter-
mixed” version of the game produced a threefold increase in players’
likelihood of associating women and science after one gameplay session.
In both of these studies, we observed little evidence of players noticing,
let alone reacting against, the game’s persuasive content because it was
not the ostensible subject or focus of the game.
An additional study involving a new version of the game for adults
(depicting workplace scenarios) and utilizing the same methodology
as the aforementioned experiment revealed the same pattern of results
with adult participants. Those participants assigned to play an “over-
Creating Stealth Game Interventions for Attitude and Behavior Change 179
loaded” version of the game exhibited significantly greater negative
affect (including the distinct response of feeling “fed up” by the end of
the game) and a lower level of concern about the issue of social biases,
compared to participants assigned to either an “intermixed” game con-
dition or a no-game control condition (Kaufman and Flanagan 2016A).
In sum, these findings confirm that over-representing serious, persua-
sive content within the game triggered players’ reactance – and that this
defensiveness prevented them from shifting their mindsets and percep-
tions after play.
In recent work, we have explored how the intermixing strategy might
also be effectively implemented in a game’s presentation of diverse
characters as a means of reducing gender bias in STEM. In the time
travel-themed strategy game The Luminists, players compete to “restore”
the most scientific and technological discoveries that have been undone
by the unraveling of time by “recruiting” real-life STEM role models
whose skills and expertise assist them in their quest. In line with prior
work demonstrating the beneficial impact of exposure to counterstereo-
typical role models for lowering social biases and increasing STEM
aspirations and pursuits (e.g., Dasgupta and Asgari 2004), the primary
underlying goal of the game was to present a host of positive female
STEM role models to young female players. At the same time, we pre-
dicted that “intermixing” female and male STEM role model “luminists”
(rather than presenting a higher ratio of female to male luminists) would
enhance the efficacy of the game – both by making the intended goals of
the game less overt and by reinforcing equity rather than imbalanced par-
ticipation in STEM between the genders. An initial experimental study
involving a sample of female youth participants supported this prediction
(Kaufman and Flanagan 2016B). In this study, we compared two ver-
sions of the game that differed in their ratio of male-to-female scientists
in the set of eight presented to players – one in which there were equal
numbers of male and female scientists and one in which six of the eight
scientists were female. Results revealed that, compared to participants
in a no-game control condition, participants assigned to play the “inter-
mixed” version of the game (but not those assigned to the “imbalanced”
180 ToDiGRA
version) exhibited significantly higher levels of psychological connec-
tion to the game’s luminists and, as a result, greater aspirations to pursue
computer programming and other STEM careers and higher self-efficacy
in STEM.
Intermixing is counterintuitive. On the surface, the strategy may seem as
though it would be less effective. Yet, despite the fact that players are
exposed to less focal content (e.g., fewer scenarios depicting occurrences
of bias in Awkward Moment or fewer female role models in The Lumin-
ists), they are significantly more likely to accept and internalize (rather
than reject and defend against) the game’s underlying persuasive aims
and messages. Our work to date has shown that the “intermixing” strat-
egy of balancing or interweaving on-topic, focal aspects of a game with
off-message or off-topic content, plays a central role in determining the
efficacy of our persuasive games.
Embedding through “Obfuscating”
The second broad Embedded Design strategy that we have employed
with great success is “obfuscating”: concealing or obscuring the true per-
suasive intent of a game by employing devices that divert players’ atten-
tion and/or allow for the covert introduction of persuasive themes or
elements. One primary example of the obfuscating method is the deci-
sion to employ a game genre whose associated goals or expectations do
not include the aim to change players’ attitudes, beliefs, or behaviors.
Indeed, our choice to design the aforementioned Awkward Moment as a
fast-paced, social party game was a wholly intentional one. Triggering
(and fulfilling) the anticipation of a fun, interactive play experience with
an abundance of levity and laughter (achieved in part through the game’s
“intermixing” of both serious and silly, or absurd Moments and Reac-
tions) created a “safe” space for players to encounter and react to the
game’s heavier, on-message content with greater comfort – and greater
candor. Our team’s extensive playtesting and iteration of the game pro-
vided consistent support; through both unobtrusive observations of play
sessions and post-game interviews with youth testers, playgroups gen-
Creating Stealth Game Interventions for Attitude and Behavior Change 181
erally approached the game with a strong and sustained spirit of levity
and amusement, yet rarely showed signs of subverting the game’s more
serious moments (or Moments), even among older, more experienced (or
even more ‘jaded’) players at venues such as the PAX or GenCon gam-
ing conventions. Moreover, even when asked directly what they believed
the true goals of the game to be, players rarely identified the game’s
primary aim of challenging gender stereotypes in STEM domains, but
rather focused more broadly on the game’s general focus on reacting to a
variety of social situations (further evidence of the successful implemen-
tation of the “intermixing” strategy) as well as a number of genre-con-
sistent goals, such as the enjoyment of the game’s social dynamics and
the amusingly random or serendipitous pairings of Moment and Reac-
tion cards that emerged.
In developing a second game with the same primary aim as Awkward
Moment – to combat stereotypes and reduce prejudices – we went even
a step further in using the party game genre to obfuscate the underlying
goals. Buffalo: The Name Dropping Game (2012) is ostensibly a rapid-
fire group trivia game: players flip a card from each of the game’s two
decks (one containing adjectives and the other nouns) and race to be the
first to shout out the name of a real or fictional person who matches
the revealed pair of descriptors (see Figure 3). What most players do
not realize (and, as playtests and interviews have revealed, are quite
surprised to learn) is that the game’s deceptively simply design was
based on an established psychological premise: exposure to a plethora
of counter-stereotypical or otherwise unexpected exemplars (to which
players are necessarily exposed given the game’s random pairings of
attributes and social categories) reliably reduces individuals’ levels of
stereotyping and prejudice. In a given play session, for example, players
may be invited to name such diverse exemplars as a “charismatic techie,”
“rugged fashion designer,” “tattooed visionary,” and “Iranian poet.”
Indeed, our own controlled experiments investigating the impact of Buf-
falo (Kaufman and Flanagan 2015) revealed that players of a single
session of the game, compared to participants in a no-game control con-
dition, exhibited significantly higher levels of social identity complexity
182 ToDiGRA
(i.e., greater diversity and inclusiveness in their perception of their pri-
mary identity groups, which is a predictor of tolerance and egalitarian-
ism: Roccas and Brewer 2002) and universal orientation (i.e., a measure
of global non-prejudice: Phillips and Ziller 1997). Thus, despite (or, we
would argue, because of) players’ general failure to realize or recognize
the game’s persuasive goals and mechanism, the game successfully shifts
players’ conceptions of their own and others’ identities simply by virtue
of playing the game and both offering and being exposed to a plethora of
exemplars of cross-cutting identity groups and associated traits. More-
over, even in cases when players recognize how their own biases might
have influenced their performance in the game (e.g., one playtest partic-
ipant regretfully reflected on his and his group’s failure to name a “His-
panic lawyer,” despite the fact that Sonia Sotomayor had recently been
appointed to the Supreme Court), they by and large do not realize that
this was, in fact, a focal outcome intended by the game’s designers.
Figure 3: Sample card pairing from Buffalo: The Name Dropping Game.
With both Awkward Moment and Buffalo, we employed yet another
means of obfuscation, one that is particularly rare among persuasive
games: we deliberately avoided disclosing the aims of the game in the
descriptions provided to players on the game box and in the instruc-
tional materials, and instead used deliberately neutral language to present
and explain the game. This choice of neutral language represents a sec-
ond obfuscation strategy: the use of framing devices that emphasize fea-
tures of the game other than its focal subject matter or persuasive aims.
Indeed, we predicted that simply revealing before play that either game
Creating Stealth Game Interventions for Attitude and Behavior Change 183
dealt in some way with social biases and stereotypes could dramati-
cally reduce players’ enjoyment of the game or the game’s impact, in
part because such “forewarning” would likely raise either conscious or
unconscious defenses in players to resist the game’s perceived intent. An
initial pair of randomized experiments (Kaufman and Flanagan 2015)
suggested that this was indeed the case. Holding all other game elements
constant, adolescent players of Awkward Moment who were randomly
assigned to a “stereotype frame” condition (and were informed prior to
play that the game dealt with “awkward social stereotypes”), compared
to those assigned to a “situation frame” condition (who were told the
game dealt with “awkward social situations”) reported finding the game
significantly less fun and immersive and failed to exhibit significantly
less movement in their rejection of gender stereotypes. Likewise, play-
groups who were told that Buffalo explored “pop culture stereotypes”
(compared to “pop culture knowledge”) did not show a reduction in
their levels of prejudice, as assessed by the measure of universal ori-
entation described above. These findings illustrate the basic premise of
the “Embedded Design” model: persuasive games that overtly telegraph
their intended purpose of shifting attitudes and mindsets are likely trig-
gering mindsets in players that hinder the game’s enjoyability and blunt
its potential positive impact.
In addition to the selection of genres and the employment of framing
language that diverts attention away from a game’s true “message,” one
final obfuscation strategy that we have applied is the delayed revela-
tion of potentially threatening, counter-attitudinal, or alienating features
or elements. Specifically, we have explored this technique to encourage
greater psychological connection and higher levels of experience-taking
with characters (Kaufman and Libby 2012): that is, greater immersion
into the role and persona of protagonists in narrative and game worlds,
particularly ones who belong to social “outgroups.” This technique has
previously proven effective for written narratives: for example, revealing
a character’s racial or sexual orientation outgroup membership later in
a short story (once a psychological connection between reader and char-
acter had begun to take root) not only facilitated higher levels of expe-
184 ToDiGRA
rience-taking but also reduced prejudice levels toward the represented
outgroups (Kaufman and Libby 2012). That is, initially obfuscating the
potentially distancing (or stigmatizing) group membership of the charac-
ter ultimately increased readers’ receptiveness of the character’s identity
– and profoundly enhanced the persuasive impact of the story.
More recently, our team successfully applied this technique to encourage
higher levels of experience-taking among male youth in our strategy
board game Monarch, which puts players in the role of sibling princesses
competing for the throne. Given the persistence of social norms that dis-
courage “gender-swapping” play, particularly among boys (e.g., Martin
1990; McCreary 1994), we anticipated that revealing their character’s
gender prior to play would reduce experience-taking among male ado-
lescent players. Conversely, in line with prior research, withholding this
revelation for several rounds (during which players became acquainted
with their characters and were gradually exposed to subtle clues about
their true identity, including the use of gowns and pageantry as political
instruments in the game) should more effectively ease players into their
cross-gender roles. This was indeed the case: a controlled experiment
(Kaufman and Flanagan 2016C) revealed that a sample of male youth
randomly assigned to play the “delayed revelation” version of the game,
compared to those assigned to an “immediate revelation” version,
reported higher levels of experience-taking with their princess characters
and, moreover, exhibited greater rejection of stereotypical gender norms
(e.g., rejection of the association between “female” and “emotional” or
“weak”) following gameplay.
Embedding through “Distancing”
The final Embedded Design strategy that we have explored in our work
is the use of “psychological distance” (Trope and Liberman 2010) to cre-
ate a safe space between individuals and the serious or sensitive themes
or topics explored or modeled by a game. By separating players from
their real-life identities and prior knowledge, beliefs, and experiences,
persuasive games can effectively circumvent players’ reticence or reluc-
Creating Stealth Game Interventions for Attitude and Behavior Change 185
tance and enhance the game’s transformative potential – particularly in
subject matter domains that may be uncomfortable or counter-attitudi-
nal. By its very nature, an absorbing, immersive game should trigger
a high level of psychological transportation (Green and Brock 2000),
thereby distancing players from their real-life surroundings and true
identities, which prior work has shown sets the stage for higher levels of
enjoyment (Green et al. 2004) and persuasion (Green and Brock 2000;
Green, Brock, and Kaufman 2004; Kaufman and Libby 2012). Indeed,
we would argue that persuasive games that take too overt or literal an
approach in their handling of controversial or sensitive topics have inher-
ently less capacity to transport their players (and to provide an enjoy-
able experience) because they create too little psychological distance to
explore those topics in a non-threatening fashion.
Beyond the psychological distance afforded by a highly transporting
game, however, there are a number of specific distancing strategies that
designers can use to increase the gap between players’ real-life expe-
riences and the ideas, encounters, and interactions that await them in
the game. These strategies are derived from a number of distinct mani-
festations of psychological distance revealed by prior work (Trope and
Liberman 2000). Perhaps the most elementary forms of distance is hypo-
theticality: rather than presenting situations that are drawn directly from
players’ real-life experiences (or situations that attempt to replicate or
mirror those experiences), encouraging players instead to engage in
“what if?” scenarios provides a safe “buffer” to explore even the most
sensitive topics. Indeed, the value of hypotheticality was a key decision
point in the design of Awkward Moment: each of the game’s Moment
Cards present a purely hypothetical situation and invites players to con-
sider a host of alternative ways of responding. Rather than placing the
players and their embarrassing moments or experiences with bias in the
spotlight, the game allows players to envision and select responses for
the unidentified, second-person “you” described in each of the game’s
A second distancing mechanism that we utilize in our work is the fic-
186 ToDiGRA
tionalizing of real-life issues and events – that is, the embedding of those
elements within symbolic, fantastical, or metaphorical representations.
This technique is by no means a new one: the use of fiction to disguise
the focus or target of a story may be as old as written language itself.
What is distinctive in our approach to using fictional representations
in games is our endeavor to systematically compare different levels of
explicitness (versus “embeddedness”) in those representations. To cite
one example of this approach, our team has designed and studied two
versions of the public health board game POX: Save the People (2011)
which is intended to promote positive attitudes and valuations toward
vaccination: one version (POX) that presented a relatively straightfor-
ward, realistic narrative about disease spread, and one version (ZOM-
BIEPOX) utilizing a more fantastical narrative about the spread of a
“zombie plague” (see Figure 4). Both games share the same essential
rules and mechanics concerning the spread of infectious disease and
the modeling of vaccination as an effective strategy for curtailing that
spread, but differ in the level of distance afforded by their representa-
tion of disease, infection, and death (or “un-death” in the case of ZOM-
BIEPOX). A pair of controlled experiments comparing the impact of
both games on both adult and youth players revealed parity between
the games (compared to a no-game control condition) in terms of their
impact on players’ valuation of vaccination as a public health solution.
At the same time, however, players of the zombie-themed version of the
game reported higher levels of psychological transportation and higher
levels of empathic concern toward individuals with infectious diseases,
as assessed by self-report measures (Kaufman, Flanagan, and Belman
2016). Thus, the use of a more distanced, metaphorical representation
of disease was not only effective in shifting attitudes toward a real-life
health policy issue but, indeed, even more effective than the less dis-
tanced, realistic narrative at forging a bond of compassion between play-
ers and the real-life individuals symbolized by the zombies in the game.
This finding lends further credence to our view that persuasive games
utilizing elements of the Embedded Design model (such as distancing)
Creating Stealth Game Interventions for Attitude and Behavior Change 187
are likely to be experienced as more transporting by players – and, con-
sequently, more impactful at changing hearts and minds.
Figure 4: POX and ZOMBIEPOX game boards.
In upcoming work, we will be exploring the benefits of fictionalized
distancing for individuals creating narrativized accounts of their own
real-life experiences – specifically, personal reflections on their expe-
riences being the target of others’ stereotypes and biased expectations,
judgments, or behaviors. This project will explore the therapeutic and
cathartic value of creating interactive “text adventures” that are based on
stressful or traumatic real-life occurrences but provide creators with the
safety (and creative license) afforded by the fictionalized re-telling and
recounting of those life events. In this stream of research we will inves-
tigate the effects of writing interactive narratives that are more fiction-
alized, versus more strictly autobiographical, as well as the impact of a
number of other distancing mechanisms, such as the narrative voice (e.g.,
a more distanced 3rd person voice versus the less distanced 1st and 2nd
person perspectives) employed by authors, on the emotional benefits of
narrativizing one’s own lived experiences.
188 ToDiGRA
The Embedded Design model offers a number of easily implementable,
evidence-based techniques that promise to revolutionize the ways that
game developers tackle serious content issues and make more effective
and more enjoyable games. As illustrated by examples from our own
game designs and accompanying empirical work, the more covert,
“stealthy” approaches derived from the Embedded Design model result
in persuasive games that are remarkably more transporting and impact-
ful, compared to games in which the message or material is presented
more overtly or directly (see Kaufman & Flanagan, 2015). Additionally,
the model advances the conversation around the application of psycho-
logical principles in games, and builds on other theoretical and practical
formulations for understanding games, such as models of game design
patterns (Bjork and Holopainen 2004).
The data emerging from empirical work on the use of Embedded Design
(via techniques such as intermixing, obfuscating, and distancing) demon-
strates that such techniques invite a more open mindset, one charac-
terized by a reduced level of activation and accessibility of players’
self-concept and predispositions, attitudes, and beliefs. Such a mindset
circumvents the psychological resistance that players are likely to expe-
rience to more overt, explicit game “interventions,” and, further, sets
the stage for players to approach and internalize new information and
ideas, take on new perspectives and roles, and understand concepts or
principles in a comparatively unbiased fashion. Indeed, the power of the
embedded approach is that it offers design solutions that have the poten-
tial to be equally effective for both individuals who may already endorse
a particular stance as well as those who may initially be opposed or indif-
ferent to it. These strategies can enable games that address social issues
to have a much broader impact.
It is important to note that the strategies described here are by no means
intended to be comprehensive. Our team has just begun to discover the
potential of such techniques. Each new game project we (and others)
Creating Stealth Game Interventions for Attitude and Behavior Change 189
take on sets the stage for new manifestations and applications of Embed-
ded Design to emerge and, as a result, extend, enhance, and refine the
design model introduced here. Moreover, although the game case stud-
ies we presented here to exemplify the model were non-digital, the prin-
ciples and practices suggested by the model are intended to be broadly
implemented across all game platforms and media delivery formats.
Indeed, the greater flexibility and control afforded by the creation of
digital games (e.g., in their revelation of information or representation
of characters) open up a world of new possibilities for embedding that
designers can consider, implement, and test in their own work (e.g., see
Christiansen 2014).
This research was supported by the U.S. National Science Foundation,
HRD-1137483 “(EAGER) Transforming STEM For Women and Girls:
Reworking Stereotypes & Bias.” Special thanks to the entire Tiltfactor
team for their contributions and practical work, which helped inform the
design model and applications offered here.
Bjork, S., and Holopainen, J. Patterns in Game Design. Charles River
Media, Newton Center, MA, 2004.
Bogost, I. Persuasive games: The expressive power of videogames. The
MIT Press, Cambridge, MA, 2007.
Brehm, J.W. A Theory of Psychological Reactance. Academic Press,
New York, 1966.
Centieiro, P., Romão, T., and Dias, A.E. “A location-based multiplayer
mobile game to encourage pro-environmental behaviours,” in Proceed-
ings of the 8th International Conference on Advances in Computer
190 ToDiGRA
Entertainment Technology (Lisbon, Portugal, November 2011), ACM
Press, p. 31.
Christiansen, P. “Presence and heuristic cues: Cognitive approaches to
persuasion in games,” in Proceedings of DiGRA 2014 (Salt Lake City,
UT, August 2014).
Dasgupta, N., and Asgari, S. “Seeing is believing: Exposure to counter-
stereotypic women leaders and its effect on the malleability of automatic
gender stereotyping,” in Journal of Experimental Social Psychology vol.
40, no. 5 (2004), pp. 642-658.
de la Hera Conde-Pumpido, T. “A conceptual model for the study of per-
suasive games,” in Proceedings of DiGRA 2013 (Atlanta, GA, August
Dunbar, N. E., Wilson, S. N., Adame, B. J., Elizondo, J., Jensen, M. L.,
Miller, C. H., … and Burgoon, J. K. “MACBETH: Development of a
training game for the mitigation of cognitive bias,” in International Jour-
nal of Game-Based Learning (IJGBL), vol. 3, no. 4 (2013), pp. 7-26.
Flanagan, M. Critical Play: Radical Game Design. The MIT Press, Cam-
bridge, MA, 2009.
Flanagan, M. and Nissenbaum, H. Values at Play in Digital Games. The
MIT Press, Cambridge, MA, 2014.
Fiellin, L.E., Hieftje, K., Duncan, L.R., Sawyer, B.G., and Fiellin, D.A.
“A videogame for substance use prevention and HIV risk reduction in
young adolescents,” in Drug and Alcohol Dependence, vol. 140 (2014),
pp. e58-e59.
Green, M.C., and Brock, T.C. “The role of transportation in the persua-
siveness of public narratives,” in Journal of Personality and Social Psy-
chology, vol. 79, no. 5 (2000), pp. 701-721.
Green, M.C., Brock, T.C., and Kaufman, G.F. “Understanding media
Creating Stealth Game Interventions for Attitude and Behavior Change 191
enjoyment: The role of transportation into narrative worlds,” in Commu-
nication Theory, vol. 14, no. 4 (2004), pp. 311-327.
Huizinga, J. Homo Ludens. Versuch einer Bestimmung des Spielele-
ments in der Kultur. Rowohlt, Hamburg, 1938.
Kaufman, G., and Flanagan, M. “A Psychologically ‘Embedded’
Approach to Designing Games for Prosocial Causes,” in Cyberpsychol-
ogy: Journal of Psychosocial Research on Cyberspace, vol. 9, no. 3
Kaufman, G. and Flanagan, M. “How Awkward! Shifting Workplace
Biases with a Social Game.” Under Review (2016A).
Kaufman, G. and Flanagan, M. “How STEM Role Models Shift Think-
ing About STEM Careers in a Time Travel Game.” Under Review
Kaufman, G., and Flanagan, M. “Not Just for Girls: A Strategy Princess
Board Game that Challenges Gender Stereotypes.” Under Review
Kaufman, G., Flanagan, M. and Belman, J. “Playing the System: Com-
paring the Efficacy and Impact of Digital and Non-digital Versions of a
Collaborative Strategy Game,Under Review (2016).
Kaufman, G. F., and Libby, L. K. “Changing beliefs and behavior
through experience-taking,” in Journal of Personality and Social Psy-
chology, vol. 103, no. 1 (2012), pp. 1-16.
Martin, C. L. “Attitudes and expectations about children with nontradi-
tional and traditional gender roles,” in Sex Roles, vol. 22, no. 3-4 (1990),
pp. 151-166.
McCreary, D.R. “The male role and avoiding femininity,” in Sex Roles,
vol. 31, no. 9-10 (1994), pp. 517-531.
192 ToDiGRA
Phillips, S. T., and Ziller, R. C. “Toward a theory and measure of the
nature of nonprejudice,” in Journal of Personality and Social Psychol-
ogy, vol. 72, no. 2 (1997), pp. 420-434.
Pronin, E., Lin, D.Y., and Ross, L. “The bias blind spot: Perceptions of
bias in self versus others,” in Personality and Social Psychology Bul-
letin, vol. 28, no. 3 (2002), pp. 369-381.
Roccas, S., and Brewer, M.B. “Social identity complexity,” in Personal-
ity and Social Psychology Review, vol. 6, no. 2 (2002), pp. 88-106.
Trope, Y., and Liberman, N. “Construal-level theory of psychological
distance,” in Psychological Review, vol. 117, no. 2 (2010), pp. 440-463.
Worchel, S., and Brehm, J. W. “Effect of threats to attitudinal freedom as
a function of agreement with the communicator,” in Journal of Personal-
ity and Social Psychology, vol. 14, no. 1 (1970), pp. 18-22.
Creating Stealth Game Interventions for Attitude and Behavior Change 193
... While also informed by work on role-play induction principles [53], the design of Chimera:Grayscale drew upon the Embedded Design Model developed by Kaufman and colleagues [32,33]. The model describes three design strategies that attempt to avoid psychological reactance (i.e., reduce the chances that a persuasive game makes the player feel threatened or defensive by being overly didactic in its approach). ...
... Empirical work does suggest that the strategies outlined by the Embedded Design Model can be effective in relation to reducing implicit biases e.g., around sexism [32,33]. However, questions remain about their potential impact on real-world behavior, and how they influence more conscious forms of reflection that the Reflective Game Design agenda [35] aims to facilitate. ...
... From an Embedded Design Model [32,33] perspective, a key design consideration is to increase the psychological distance [50] between the player and the game so that players can explore sensitive topics in a non-threatening way. Based on prior work indicating that higher levels of psychological transportation lead to higher levels of enjoyment [26] and persuasion [25,26,34], the goal is to create "an absorbing, immersive game [that] should trigger a high level of psychological transportation thereby distancing players from their real-life surroundings and true identities" (p. ...
Persuasive games promote attitude and behavior change, of which reflection is an important precursor, but existing advice on designing for reflection is mixed and requires further empirical investigation. To address these concerns, we report on the design and evaluation (n=32) of a game to prompt student reflection on work-life balance. Participants either played as themselves or a third person character (Alex). An inductive qualitative analysis of post-play interviews, and a follow-up one week later, resulted in four themes that consider how gameplay facilitated reflection: making (sensible) consequences visible; it’s like MY life; the space between Alex and I; and triggers in everyday life. In addition, a deductive qualitative analysis indicated that while both games resulted in different forms of reflection for the majority of players, those who role-played as Alex appeared more likely to experience higher levels of reflection. Through exploring the different ways that the two versions of the game succeeded, and failed, to support reflection, we highlight the importance of providing a relevant context to players (so the game feels close to their experience), and allowing them to role-play as someone other than themselves (but not too close).
... Esta carencia de sutileza al enviar el mensaje hace que los consumidores lo reciban con las defensas altas, críticamente, conscientes de que se les intenta persuadir, lo cual disminuye su efectividad (Kaufman et al., 2015). Por ello, cuando se han realizado estudios empíricos para comprobar la eficacia empírica de estos dispositivos, sus resultados son más bien limitados (Soekarjo, 2015). ...
... A través de la experiencia del/de la jugador/a, determinados valores son normalizados, de igual modo que los juegos persuasivos esperan que sus mecánicas hagan que el/la jugador/a cambie de opinión. Pero lo hacen de un modo implícito, por ello más eficaz ya que actúa de modo sutil (Kaufman et al., 2015). ...
Full-text available
Los videojuegos son una de las industrias culturales más pujantes que existen en la actualidad, lo cual los convierte en un altavoz privilegiado para toda suerte de discursos. Sus mensajes también han ganado en madurez y profundidad, y se han convertido en un entorno usado para la persuasión y la creación de identidades que permean con intensidad en el ámbito privado de millones de jugadores y jugadoras. En este trabajo estudiamos las mediaciones y los procesos de cocreación de identidades en los videojuegos, bajo las lógicas del capitalismo tardío, y cómo dicho rasgo está presente de manera ubicua en cualquier género videolúdico más allá de los denominados juegos persuasivos.
... Hodson et al. (2009) assessed attitude change toward homosexuals using simulation intervention, in which participants imagined life on an alien planet with constraints parallel to those faced by homosexuals. Kaufman et al. (2016) created a stealth game intervention to increase awareness of critical and timely social issues to change attitudes (and behaviors) through gameplay. Rosler et al. (2022) tested how Informative Process Model-based videos may encourage conflict-supporting narratives. ...
Full-text available
According to the Dialogical Self Theory, there is a multiplicity of I-positions within the self in a dialogical relationship. Psychodrama refers to these positions as inner parts within the self. We propose an integration of both to create an intervention centered on unscripted intrapersonal role-playing of the I-positions for the purpose of meaning-making. We want less dominant or silenced I-positions to be heard and included in the dialogue. We illustrate this approach with a group of 24 adult participants who engaged in a reflective dialogue process to potentially change their attitudes and perspectives toward an exemplary topic of technological advancement. A mixed-method approach is applied to investigate the potential benefits of this intervention. In pre-and post-intervention, a questionnaire is used to gather self-reported attitudes with Likert-scale and open-ended questions to assess explicit attitudes and perspectives related to our exemplary case. We used a statistical test and thematic analysis to compare the results before and after the intervention. The qualitative results show a shift in perspectives. However, the quantitative results do not show statistical significance regarding attitude change. Furthermore, integrating quantitative and qualitative results shows a convergence of the findings in one group of participants while divergence in the other. Despite these results, the proposed interventional approach promises various potential applications.
... It is likely to trigger psychological defenses within the user that can block the persuasive aim or even inspire beliefs that are contrary to the intended outcomes. Embedded design is instead based on the premise that hiding educational social content within the game makes the player more receptive to the learning process [60]. ...
Full-text available
Serious games have to meet certain characteristics relating to gameplay and educational content to be effective as educational tools. There are some models that evaluate these aspects, but they usually lack a good balance between both ludic and learning requirements, and provide no guide for the design of new games. This study develops the Gaming Educational Balanced (GEB)Model which addresses these two limitations. GEB is based on the Mechanics, Dynamics and Aesthetics framework and the Four Pillars of Educational Games theory. This model defines a metric to evaluate serious games, which can also be followed to guide their subsequent development. This rubric is tested with three indie serious games developed using different genres to raise awareness of mental illnesses. This evaluation revealed two main issues: the three games returned good results for gameplay, but the application of educational content was deficient, due in all likelihood to the lack of expert educators participating in their development. A statistical and machine learning validation of the results is also performed to ensure that the GEB metric features are clearly explained and the players are able to evaluate them correctly. These results underline the usefulness of the new metric tool for identifying game design strengths and weaknesses. Future works will apply this metric to more serious games to further test its effectiveness and to guide the design of new serious games.
... These findings provide an important foundation for research into whether Flex and future related intervention strategies may provide effective means to intervene with exercise behavior, without relying on persuasion or education -without targeting intentions. Such approaches provide possibilities of 'stealth' physical activity interventions (Kaufman et al., 2016) by engaging users in valued activities (e.g., app game play) with the intended side-effect of altering motivation for promoting another behavior (e.g., physical activity). ...
... Our reasons for doing so were two-fold: first, as explained above, the current political environment, especially around the term "fake news" as a political rallying cry, meant that utilizing real news or current events as the basis for the game could create undesirable tension between players and be a distraction from the key learning goals. The sheep setting created psychological distance to allow players to engage with the concept of fake news in a less polarizing context, a technique with demonstrated effectiveness in persuasive games (Kaufman et al., 2016). Second, using fictional news stories freed us from the constraints of there being a set correct answer regarding whether any given story was "real" or "fake," allowing that determination to be made by the players within the game, using their available evidence each round. ...
Full-text available
Given the need for innovative, engaging, and youth-centered approaches to media literacy, as well as the potential of active pedagogies to facilitate youth civic education and efficacy, games emerge as a particularly promising and under-utilized avenue for news literacy education. Our research asks, how might we use game-based learning to tackle fake news and stimulate news literacy among a youth audience? Here, we reflect on the process of designing LAMBOOZLED!, a news literacy game for middle school and high school students, based on a multilevel game design framework that allowed us to articulate learning objectives, consider suitable mechanics, dynamics and aesthetics, and integrate relevant instructional principles along multiple learning dimensions. Positioning this work at the nexus of game design and media literacy education, we discuss our key decision points and the larger stakes of adopting a game-based approach to news literacy education in the current political climate.
... However many PD approaches are limited in their ability to facilitate generative conversation around sensitive subjects such as racism, particularly with groups of strangers. Our approach leverages the psychologically protective benefts of fction (e.g., [15,55,61]) as well as a social justice and critical race lens to build on the generative and human-centered aspects of participatory design. ...
... 1.2.2), any study showing the effects of 'the good' also could propagate 'the bad'. I focus on the way games can set players to elaborate on a game-based message to come to their own conclusions, rather than looking at more subversive, stealthy processes (e.g. as discussed by Kaufman, Flanagan, & Seidman, 2016). ...
... Indeed, addressing sensitive topics frequently involves seeking for transgressive waysgameplays-of dealing with moral arguments. In particular, we focused on topics of questionable playful interest, usually steeped in negative value and avoided in daily conversations (Kaufman, Flanagan & Seidman, 2016): we asked our students to explore the game and its dark side (Mortensen, Linderoth & Brown, 2015), deliberately investigating morally defensible choices and giving rise to negative emotions and feelings (Juul, 2013). Designing, prototyping and testing these games prompted rumination between meanings shared and transferred through the gameplay and the experience they trigger (Sicart, 2011a). ...
Responses to urban human services issues such as housing and unemployment often overlook lived experiences through these systems and are formulated from a top-down (systems, services, or policy-level) perspective. This study integrates systems thinking and design justice principles for centering the voices of those experiencing these issues towards exploring ‘agency'—the capacity to act—from the bottom-up and top-down in responding to these issues. An agency typology encompassing various bottom-up and top-down agencies is developed through an analysis of Digital Games for Change (DG4C) for the various agencies they can initiate. The agency typology's contributions are threefold—in research (as a method and analytical tool), in practice (as design principles) and in education (for teaching collective action, impact). The agency typology can drive ‘concerted agency' or collective action, where top-down and bottom-up agencies work together, enabling multipronged targeted approaches to complex social issues and maximizing social justice efforts through collective impact.
Full-text available
Prosocial games often utilize a direct, explicit approach to engage players with serious real-life scenarios and present information about key societal issues. This approach, however, may limit a game’s persuasive impact and ability to produce beneficial outcomes, particularly when the apparent aims of the game trigger players’ psychological defenses or reduce players’ potential engagement with-and enjoyment of-the game experience. In contrast, the “Embedded Design” approach that we introduce here offers effective, evidence-based strategies for more stealthily or covertly delivering persuasive content in a game in a fashion that circumvents players’ psychological defenses and triggers a more receptive mindset. This paper provides an in-depth exploration of two key Embedded Design strategies: (1) intermixing: combining “on-topic” and “off-topic” game content in order to make the focal message or theme less obvious and more accessible and (2) obfuscating: using game genres or framing devices that direct players’ attention or expectations away from the game’s true aims. To illustrate the implementation and effectiveness of these strategies, we detail the design of two games that utilize a number of these techniques to reduce stereotypes and biases and present the methods and results of a set of empirical studies testing the prosocial impact of these games. In addition, we introduce a number of other Embedded Design strategies that have emerged in our work and discuss the most viable contexts for the use of this design approach. © 2007-2015 Cyberpsychology: Journal of Psychosocial Research on Cyberspace.
Full-text available
This paper describes the process of rapid iterative prototyping used by a research team developing a training video game for the Sirius program funded by the Intelligence Advanced Research Projects Activity (IARPA). Described are three stages of development, including a paper prototype, and builds for alpha and beta testing. Game development is documented, and the process of playtesting is reviewed with a focus on the challenges and lessons-learned. Advances made in the development of the game through the playtesting process are discussed along with implications of the rapid iterative prototyping approach.
In this article, we introduce the concept of social identity complexity—a new theoretical construct that refers to an individual's subjective representation of the interrelationships among his or her multiple group identities. Social identity complexity reflects the degree of overlap perceived to exist between groups of which a person is simultaneously a member When the overlap of multiple ingroups is perceived to be high, the individual maintains a relatively simplified identity structure whereby memberships in different groups converge to form a single ingroup identification. When a person acknowledges, and accepts, that memberships in multiple ingroups are not fully convergent or overlapping, the associated identity structure is both more inclusive and more complex. In this article, we define the concept of social identity complexity and discuss its possible antecedents and consequences. Results from initial studies support the prediction that social identity complexity is affected by stress and is related to personal value priorities and to tolerance of outgroup members.
Three studies suggest that individuals see the existence and operation of cognitive and motivational biases much more in others than in themselves. Study 1 provides evidence from three surveys that people rate themselves as less subject to various biases than the “average American,” classmates in a seminar, and fellow airport travelers. Data from the third survey further suggest that such claims arise from the interplay among availability biases and self-enhancement motives. Participants in one follow-up study who showed the better-than-average bias insisted that their self-assessments were accurate and objective even after reading a description of how they could have been affected by the relevant bias. Participants in a final study reported their peer’s self-serving attributions regarding test performance to be biased but their own similarly self-serving attributions to be free of bias. The relevance of these phenomena to naïve realism and to conflict, misunderstanding, and dispute resolution is discussed.