Graduate Theses and Dissertations
Iowa State University Capstones, Theses and
Dissertations
2020
The will and the skill: The training effects of virtual reality and
The will and the skill: The training effects of virtual reality and
gaming
gaming
Andreas Miles-Novelo
Iowa State University
Follow this and additional works at: https://lib.dr.iastate.edu/etd
Recommended Citation
Recommended Citation
Miles-Novelo, Andreas, “The will and the skill: The training effects of virtual reality and gaming” (2020).
Graduate Theses and Dissertations. 18079.
https://lib.dr.iastate.edu/etd/18079
This Thesis is brought to you for free and open access by the Iowa State University Capstones, Theses and
Dissertations at Iowa State University Digital Repository. It has been accepted for inclusion in Graduate Theses and
Dissertations by an authorized administrator of Iowa State University Digital Repository. For more information,
please contact digirep@iastate.edu.
The will and the skill: The training effects of virtual reality and gaming
by
Andreas Miles-Novelo
A thesis submitted to the graduate faculty
in partial fulfillment of the requirements for the degree of
MASTER OF SCIENCE
Co-Majors: Psychology; Human-Computer Interaction
Program of Study Committee:
Craig A. Anderson, Major Professor
Jonathan Kelly
Douglas Gentile
The student author, whose presentation of the scholarship herein was approved by the program
of study committee, is solely responsible for the content of this thesis. The Graduate College will
ensure this thesis is globally accessible and will not permit alterations after a degree is conferred.
Iowa State University
Ames, Iowa
2020
Copyright © Andreas Miles-Novelo, 2020. All rights reserved.
ii
TABLE OF CONTENTS
ABSTRACT
…………………………………………………………………………………………………………………. IV
CHAPTER 1. GENERAL INTRODUCTION ……………………………………………………………………. 1
Gaming, Virtual Reality, and Training ……………………………………………………………………. 1
Gaming Research – A History ……………………………………………………………………………….. 1
Virtual Reality
……………………………………………………………………………………………………… 3
Virtual Reality and Game Training
…………………………………………………………………………. 5
Efficacy, Confidence, and Willingness to Engage in a Task ………………………………………. 8
Present Research ………………………………………………………………………………………………….. 9
CHAPTER 2. METHODS ……………………………………………………………………………………………… 11
Design ………………………………………………………………………………………………………………. 11
Participants
………………………………………………………………………………………………………… 11
Drop-Outs and Missing Data ………………………………………………………………………….. 11
Demographics ……………………………………………………………………………………………………. 12
Materials …………………………………………………………………………………………………………… 13
Skill Testing …………………………………………………………………………………………………. 13
Games …………………………………………………………………………………………………………. 14
Self-Reported Efficacy
…………………………………………………………………………………… 16
Control Variables and Suspicion …………………………………………………………………….. 17
Procedure ………………………………………………………………………………………………………….. 18
Hypotheses
………………………………………………………………………………………………………… 20
Analysis Plan …………………………………………………………………………………………………….. 21
CHAPTER 3. RESULTS ……………………………………………………………………………………………….. 23
Sample Size and Power
……………………………………………………………………………………….. 23
Outliers
……………………………………………………………………………………………………………… 24
Correlations of Variables of Interest
……………………………………………………………………… 24
Hypothesis Testing
……………………………………………………………………………………………… 26
Preliminary Analyses …………………………………………………………………………………….. 26
Main Analyses: Z-Scores and Residuals
…………………………………………………………… 27
Improvement of Performance …………………………………………………………………………. 30
Change of Self-Efficacy
…………………………………………………………………………………. 32
CHAPTER 4. GENERAL DISCUSSION AND CONCLUSION
………………………………………… 36
REFERENCES …………………………………………………………………………………………………………….. 43
iii
APPENDIX A. PAPER TARGET DISTANCE
………………………………………………………………… 49
APPENDIX B. AIRSOFT GUN
……………………………………………………………………………………… 50
APPENDIX C. PAPER TARGET
…………………………………………………………………………………… 51
APPENDIX D. PUTTING MATT ………………………………………………………………………………….. 52
APPENDIX E. SCREENSHOT FROM BATTLEFIELD 4
………………………………………………… 53
APPENDIX F. SCREENSHOT FROM THE GUN CLUB VR
…………………………………………… 54
APPENDIX G. VIVE CONTROLLERS
………………………………………………………………………….. 55
APPENDIX H. THE EFFICACY OF SHOOTING BEHAVIORS SCALE
………………………….. 56
APPENDIX I. GOLF EFFICACY SCALE ………………………………………………………………………. 58
APPENDIX J. PRE-TRAINING SURVEY ……………………………………………………………………… 59
APPENDIX K. AGGRESSIVENESS SCALE
………………………………………………………………….. 62
APPENDIX L. POST- TRAINING SESSION SURVEYS ………………………………………………… 64
APPENDIX M. INFORMED CONSENT FORM
……………………………………………………………… 68
APPENDIX N. GUN TEST SCORES BOXPLOTS
………………………………………………………….. 71
APPENDIX O. GOLF TEST SCORES BOXPLOTS ………………………………………………………… 72
APPENDIX P. GUN AND GOLF EFFICACY SCORES BOXPLOTS
……………………………….. 73
APPENDIX Q. IRB APPROVAL …………………………………………………………………………………… 74
iv
ABSTRACT
The effects of video games and virtual simulations have long been researched, and we
know that engaging in these games and situations can have a multitude of effects. Video games
and simulations have been shown as effective learning tools for people training to become
surgeons (Duque, Fung, Mallet, Posel, & Fleiszer, 2008; Rosser et al., 2007; Seymour et al., 2002),
fly planes (Dennis & Harris, 1998), drive (Ivancic & Hesketh, 2010), for military training
(Williamson et al., 2011; Curry et al., 2016), golf practice (Pohira-Vieth, 2010), and can even have
other uses such as training for physical therapy (Betker, Desai, Nett, Kapadia, & Szturm, 2007).
However, do these effects translate to everyday gaming habits, and could they teach participants
to shoot a gun or putt a golf ball? Moreover, are these effects altered at all by the introduction of
virtual reality? We also know that willingness to engage with a task is determined by familiarity
with the task, attitudes towards the task, and confidence in completing the task (also known as
self-efficacy; Bandura, 1977b). This leads us to ask, does exposure to and practice of a skill in
virtual environments (such as video games) increase one’s self-efficacy? Discussed is a study
where 100 participants engaged in a training module to model practicing shooting a gun or putting
a golf ball, and that measured whether exposure to practicing these skills in a virtual environment
(versus a 2D gaming environment) lead to greater reports of self-efficacy on those tasks. Results
found that while practicing putting or shooting showed increases in specific task performance and
self-efficacy, these were not moderated by how participants practiced these skills (in real-life, VR,
VR, by a traditional video game setup). However, these results need further research due to
concerns such as statistical power when looking at interactions.
1
CHAPTER 1. GENERAL INTRODUCTION
Gaming, Virtual Reality, and Training
The end of the twentieth century and the onset of the twenty-first has brought
about scientific and technological changes beyond comprehension and at a rate almost
cumbersome to keep up with. From computers, cell phones, the internet, and even
mailable groceries, our lives are changing drastically every day with each technological
improvement.
For years, the forefront of new technology has been in the realm of video games.
From putting the first computers (at least computer chips) in people’s homes to
innovating interactive UI’s, video games have consistently been a driving force behind
technological advances, and even more subtly, our consumption of them. While
psychological research (from all fields) has been done on the effects of video games and
even technological use on our brains and behavior, we have some fundamental questions
left unanswered.
Gaming Research – A History
We know video games are effective learning tools, as they give us unique ways of
providing feedback and motivation (Gentile & Gentile, 2008). They even fit more
classical models of learning, such as giving us role models to help us learn new behaviors
(Bandura, 1977a). Observational learning is a powerful teaching technique, and one that
occurs often. In fact, observational learning is so powerful that not only do we see it in
groups such as kids (Bandura, 1977a; Wyre, 2017), but in other species such as octopi
(Fiorito & Scotto, 2018), chimpanzees (Tomasello, Davis-Dasilva, Camak, & Bard,
1983), and rats (Heyes & Dawson, 1990).
2
Games are unique in our consumption of them due to their inherent interactive
component. They offer consumers a more “active” approach than traditional visual or
auditory media (books, movies, television) and research has shown us that learning is
best when the learner is “active” (Atlas, Cornett, Lane, & Napier, 1997).
Research in the area of game learning is fairly substantial as well. We know that
games can help increase a variety of training, including but not limited to, things such as
surgery (Duque, Fung, Mallet, Posel, & Fleiszer, 2008; Seymour et al., 2002; Rosser,
Lynch, Haskamp, Gentile, & Yalif, 2007), flight (Dennis & Harris, 2009; Gopher, Weil,
& Bareket, 1994), prosocial skills (Flynn, Palma, & Bender, 2007), military training
(Williamson et al., 2011; Curry et al., 2016), golf (Pohira-Vieth, 2010), and can even help
teach physical therapy techniques (Betker et al., 2007; Flynn et al., 2007).
Video games have also (famously) been known to increase things such as hand-
eye coordination, and visuospatial awareness (Green & Bavelier, 2003; Griffith,
Voloschin, Gibb, & Bailey, 1983). Because of these, we assume that some of these
training effects must carry into other areas.
However, as with any societal or technological advances, both positive and
negative effects arise. Amongst the examples listed, concern about violent video games
has increased, especially as they have continued to see innovation leading them to
become faster, more accessible, and more realistic. The concern not only extends to their
overall appearance but their realism in controls and elicited emotions as well. To quote a
recent YouTube video, “Call of Duty can only get more real if it gives you actual PTSD”
(Patrick, 2017). It is well documented that playing violent video games leads to increased
effects in aggression, even cross-culturally (Anderson et al., 2011; 2017), and some
3
researchers have gone so far to make declarations that some of these games can act as,
“murder simulators” (Grossman & DeGaetano, 1999).
While there has been significant debate about the effects of violent video
gameplay on aggression, meta-analyses (Anderson et al., 2010; Greitemeyer & Mügge,
2014) shed light not only on if the effect is real (it most certainly is) but the pervasiveness
of the effect, and the soundness of the science behind it. Anderson et al. (2010) clearly
show (even across study design, and even across cultures) that violent video game play
leads to (and predicts) increased aggressive behavior, aggressive affect, and aggressive
cognition, as well as decreased empathy and prosocial behavior in participants.
With the consistent improvement and access to better graphical fidelity, and more
realistic and responsive controls, the question now becomes if the players are absorbing
any of the skills portrayed in these games, and if they are, how we could use this
technology to train people, and what sort of effects this is having on the general
population?
Virtual Reality
As noted earlier, gaming has become a unique art form and mode of entertainment
because of its interactive nature. As technology has improved, so has our ability to
interact with it. We also know that this interactivity facilitates identification with the
characters (another predictor for learning) even if they are violent in portrayal (Konijn,
Bijvank, & Bushman, 2007).
At the onset of the twenty-first century, we saw an increased rise in motion
controls (the ability to control an electronic interface using our physical motion).
Previous researchers have found that a greater transfer of skills occurs when the
4
controller is more realistic (Gentile, 2011; Pohira-Vieth, 2010; McGloin et al., 2015;
Whitaker & Bushman, 2012), and motion controls have allowed some basic models for
this (controllers that look like golf clubs, guns, swords, steering wheels, etc.). There is
evidence suggesting that simple observational learning occurs in more traditional gaming
contexts with a controller or mouse and keyboard as well (Bushman, 2018). In fact, the
British military has used “commercial” gaming technologies to train their soldiers for
several years (Curry et al., 2016).
The improvements for more accurate, and cheaper motion controls, as well as
other revolutions in motion capture technology, have led the way to innovations in other
technologies once thought as distant fantasy, such as virtual and augmented reality. While
the dystopian views of these concepts found in popular culture (such as Ernest Cline’s
novel Ready Player One, and popular films like The Matrix) are likely still far off from
being possibilities, the first wave of consumer-friendly versions of this technology have
entered the household.
Recently, we have seen the wide release of the “big three” video game virtual
reality head-mounted displays: HTC’s Vive, the Oculus Rift, and Sony’s PlayStation VR
(the latter of which has made full-scale VR giving the cheapest it has ever been for
consumers). Within the last couple of years as well, has also seen many name-band
technology companies push out “cheaper” virtual reality headsets, such as Samsung’s
Gear VR and Google Cardboard, which are compatible with smartphones.
Augmented reality has seen massive and similar innovation as well. Three years
ago, at the Electronic Entertainment Expo (E3), Microsoft announced their plans to
support augmented reality for their home video game consoles, known as the
5
“HoloLens.” Originally a consumer version was originally slated for winter of 2019,
though Microsoft has instead shifted its focus for the device on more industry
applications. However, augmented reality has become standard technology as games such
as Pokémon Go! have normalized the use of AR for cellular devices. Companies such as
Ikea have begun introducing AR apps as well to help consumers picture what a piece of
furniture will look like in their home, and the wine brand 19 Crimes allows you to
interact with the artwork on their bottles.
With these technologies already making their way into our homes, research has
begun to look into the effects and uses of them. However, one thing to note is that we are
still in what one could call the “Pong phase” of these technologies. While yes, the video
game Pong was revolutionary, no one was imagining that the virtual game of paddle ball
would lead to something such as Call of Duty: World War II, where it’s opening moments
feel as if you are suddenly in control of the film Saving Private Ryan. What we can do in
the space of VR hasn’t been fully realized, so the importance of understanding its basic
effects are important as the development of the technology continues.
Virtual Reality and Game Training
Some preliminary research on training in VR has been done with varying effects.
The military has already begun trying to construct virtual environments to train soldiers
and have been able to construct realistic environments using hardware and software
commercially available from the major video game companies (Williamson et al., 2011).
In 2002, researchers found that virtual reality training increased skills in the performance
of laparoscopic cholecystectomy and found that students who trained to do the surgery
using a virtual reality program performed just as well as students who trained more
6
classically (Seymour., Gallagher, Roman, O’Brien, Bansal, Andersen, & Satava, 2002).
These results have been tested and retested numerous times (Grancharov, Kristiansen,
Bendix, Bardram, Rosenberg, & Funch-Jensen, 2003) to find similar results. However, a
common issue in the medical literature is that these samples tend to be small (N=16 in
both of the studies mentioned), so further research into the effects has yet to fully been
investigated.
For a more extreme example of virtual training, in 2014, Jann Mardenborough
entered a competition playing Gran Turismo (a popular driving simulator game) where
the winner had a chance to join a professional driving team. Not only did Mr.
Mardenborogouh win the competition, but he also qualified for the team, and is making a
career of being in professional driving. In a 2014 interview, Mardenborough said, “The
game is quite similar to real-life…in GT6 you can really feel the car move underneath
you, and if you put too much yaw or pitch into the car, it will react as it would in real life
with a bit of a snap. It was crazy to jump from the virtual world into a sports car for the
first time. I didn’t feel like there was much difference in terms of the way the car feels
and how you deliver your inputs” (Barron, 2013).
These findings and examples give us a theoretical and practical direction of
inquiry, as it seems to be evident that gaming, and virtual reality, can be used for training
different skills. In Grancharov et al. (2003), the participants were also given a baseline
psychomotor abilities test, and the results showed that those who trained in the VR
program were 29% faster at the gallbladder surgery and were less likely to commit errors
than those who did not receive the VR training. There also are practical reasons to pursue
this as well. There is more room for error and repetition in virtual training than traditional
7
training, materials can cost less (in the long run), and there can be more objective and
efficient forms of feedback.
The question of if these sorts of effects could translate to gun use have yet to be
fully researched as well. With the effects of aggression known from video games, the
question then arises, are we not only exposing players to a great risk factor of aggression
but also giving them the training to act out this aggression?
There are also practical questions to be asked here as well. Additionally, as
technology continues to become more refined, it will become less expensive. If we can
have similar, or even better, training effects from virtual training, this could potentially
save time, money, and decrease the risk of training people in using firearms. This could
be particularly useful for our military, who have already begun training soldiers using
video games (Grossman, 1995).
In 2018, Brad Bushman looked into the question of if games could train people to
shoot a gun, and found that participants who played a violent shooting game were more
likely to shoot at a mannequin’s head if the game reinforced higher rewards for getting
headshots, and this effect was true regardless of if participants used a normal video game
controller or one shaped like a pistol. The purposes of the present study are to take the
paradigms from Bushman (2018) but to repurpose the pistol-shaped gun with virtual
reality.
We also hope to take these learning and training effects to a more generalizable
level and want to see if we can find a similar effect by looking at another skill, putting a
golf ball.
8
Furthermore, when we ask these questions in light of the aggression research,
especially the violent media effect, we can’t help but ask, are we enabling people to have
the means to act aggressively? To feel competent when they act aggressively? To have
the confidence to act aggressively? These questions also become pertinent when one
exposed to many risk factors may now be given the confidence to act aggressively
through their potential virtual training.
Efficacy, Confidence, and Willingness to Engage in a Task
Classic research into efficacy expectations tells us that when people feel like they
can perform a certain task well, they will be more likely to engage in it, and in general, as
knowledge and comfort with a task increase, so do the likelihood of engaging in it. In
fact, the more confident we are that we will perform well, the more likely it is that we
will perform well (Lent, Brown, & Larkin, 1984; Bentz & Hackett, 1983). We also know
that merely imagining oneself engaging in a behavior, such as donating blood, can
increase one’s expectations and behavioral commitments to participating in that behavior
(Anderson, 1983; Anderson & Godfry, 1987).
In 1977, Bandura posited his theory of self-efficacy (1997b). Self-efficacy is
defined as a person’s belief in their ability to perform a specific behavior needed to
influence specific outcomes, “whereas confidence is the feeling or consciousness of one’s
powers or reliance on one’s circumstances” (Finch, Weiley, Edward, & Barkin, 2008).
Bandura more specifically defines the difference between efficacy expectations (the
belief one can perform the intended behavior) from outcome expectations (believing that
the behavior will have the expected outcome). So, one can feel confident that they can
perform a task or behavior, but not feel as if they can control the outcome of the situation.
9
Finch, Weiley, Edward, & Barkin (2008) looked at how perceived self-efficacy
influenced confidence in violence prevention counseling. What they found was that
pediatricians were more confident in discussing violence prevention topics such as
limiting media violence and different discipline technique, but when it came to other
topics such as gun storage and removal, they were less confident and less likely to discuss
them. They also note, “that low perceived self-efficacy and lack of self-confidence are
deterrents to delivery of prevention services, such as dental screening, screening for risky
health behaviors, domestic violence screening, and safety and family issues (Dela Cruz,
Rozier, & Slade, 2004; Ozer, Adams, Gardner, Mailloux, Wibbelsman, & Irwin, 2004).”
Another example stems from Cheng et al. (1999), which found that 93% of
pediatricians had a high level of confidence in discussing the potential dangers of second-
hand smoke. However, only 56% believed that they could do anything to prevent it.
Likewise, Maiuro et al. (2000) identified perceived self-efficacy as a moderator in
discussing domestic violence during domestic violence counseling sessions.
We hypothesize that not only do games have a training effect on actual skill (as
discussed earlier) but that they will make participants feel more comfortable and
confident in the tasks being portrayed on the screen and being enacted by the player, thus
making the player more likely to engage in them in a real-world scenario.
Present Research
In this study, participants were evaluated on their skills in shooting an airsoft gun
at a paper target, and at putting a golf ball on an office putting green. They were
randomly assigned to be trained in one of six conditions (see table 1) over a 2 to 3-week
period.
10
Participants then completed six sessions total over this period. The first session
included a baseline assessment of their skills in putting and shooting, as well as
introductory surveys and assessments of attitudes. The next four sessions were the
“training” sessions, in which they practiced the (one) skill randomly assigned to them for
twenty-minute sessions, as well as completed some additional surveys. The final session
was a post-training test, where participants came back to the lab and tested again on their
shooting and putting skills, as well as retook the various attitudes and efficacy scales they
initially completed.
11
CHAPTER 2. METHODS
Design
The study was a 2 (Skill practiced: golf or gun) by 3 (Practice Mode: VR game,
“flat screen” game, or “practical” training) by 2 (Skill measured: shooting or putting)
mixed design, with the first two factors being between participants and the third factor
being a within-participants factor. The dependent variables were scores on attitudes and
efficacy beliefs concerning golf and guns and accuracy on shooting and putting. To better
assess (and control) gender effects, participants were blocked by gender and then
randomly assigned to one of the six training conditions. Participants took pre and post-
tests on accuracy for both shooting and putting (e.g., a person who practiced the VR gun
condition still took the putting test). This allows for the scores on the skills participants
did not train on, to act as a de-facto control group.
Table 1
List of Six Conditions
Skill Practiced
Shooting
Putting
VR
VVR
Practice Mode
Flat Screen/Controller
Flat Screen/Controller
Practical
Practical
Participants
Drop-Outs and Missing Data
One hundred thirty-eight undergraduate participants from a large midwestern
university participated in the study and were given course credit for their participation.
12
However, some participants failed to complete the study or took too long to complete all
six sessions. Those who strayed outside of the targeted time frame (21-28 days to
completion), did not complete the second testing session, or experienced experimenter
error, were removed from data analysis, this totaled to 38 participants. Potential future
analyses could include some of these participants using missing data procedures for
longitudinal analyses and missing time points (e.g., Twisk & Vente, 2002) to
accommodate for these issues in drop-out. Means were calculated to supplant any
missing data from the final 100 participants included.
Demographics
For the final analysis, 100 participants were analyzed after using the above
exclusion criteria: 60 males (60%) and 40 females (40%). The average participant age
was 19.21 years old (SD = 1.55), and most identified as being Caucasian (76%). 14
(14%) participants completed the VR shooting condition, 14 (14%) completed the VR
golf condition, 15 (15%) completed the practical (“real-life”) gun condition, 17 (17%)
completed the practical golfing condition, and finally, 20 (20%) participants completed
the “controller” golf and shooting video game condition respectively. Additionally, 49
participants reported having used a gun prior to the study (49%; M (of gun use, 1-4 scale,
see materials) = .85, SD = 1.12) and 66 reported having gone golfing before (66%; M =
1.13, SD = 1.02). Only 19 participants reported spending no time playing video games
weekly (19%; M = 2.28, SD = .87).
13
Materials
Skill Testing
For this study, participants were evaluated by performance on shooting a target
with an airsoft gun, and how well they perform at putting a golf ball (accuracy of shots,
number of shots). Participants were brought into the lab and randomly assigned to a golf
or gun training condition. They then would take a baseline measure of performance for
both the shooting task and the golf-putting task.
Participants used a battery-powered airsoft gun to shoot 16 “bullets” (airsoft
pellets) at a paper target approximately 10-12 feet away from them (see appendices a-b).
Research assistants (RAs) recorded how many shots hit the target, and how many missed.
The “score” was then calculated by measuring the distance how close each of the shots
was to the “bullseye.” A hit on the bullseye was scored “zero,” and shots outside of that
were measured with measuring tape and recorded to the quarter of the inch. The closer a
participant’s score was to zero, the more “accurate” that participant was in this task. If
participants missed the paper target all-together, the score would be marked as a “9.”
Participants were also given 16 shots with a putter on an “office” golf putting set
that was roughly 9 feet long (see appendix d). After each “stroke” from the set location,
RAs measured how far away the ball was from the hole. Participants putted at a marked
spot nine away from their target hole for each of their 16 shots. Like the shooting task,
this was a “zero-score” measure, where if participants hit the ball in the hole, they
received a score of zero, and if they missed the hole, the distance to the hole was
measured to the quarter of the inch. Scores closer to zero were considered the most
“accurate.”
14
These assessments were given as a pre and post-test measure to see if partaking in
the training conditions helped participants improve their scores from test one to test two.
As stated before, participants completed both tests, regardless of which training condition
they engage in to act as a control measure. These scores were calculated by summing the
total distance from the targets.
Games
Four different games were used for the different conditions in this study. The
game used for the “traditional game” shooting range was Battlefield 4, which features a
shooting range area that players can practice shooting stationary targets (see appendix e).
For the virtual reality game, participants played Gun Club VR, which uses the Vive’s
motion controllers as the primary form of controls (see appendix f). For both the
traditional golf game and the VR golf game (which was also played on the Vive using its
touch controllers – see appendix g), players played The Golf Club which has both
traditional and (see appendix h) virtual reality modes, as well as a mode to practice
putting specifically.
For the shooting range VR game, participants stood in the middle of the lab space
with the VR headset on, while holding a controller in their hand. Presented to them on the
screen would be a shooting range, where they would then practice shooting with a
handgun for 20 minutes. The perspective in the headset was in first-person, and the one
would use the Vive controller like you would a gun, bringing it to your face to look down
the scope, and pulling the trigger on the back of the controller to shoot (much like pulling
the trigger of a gun).
15
For the shooting range controller game, participants would load into a shooting
range in Battlefield 4. Just like the VR game, the point of view was first person, and on
the screen, participants could see their hands holding the gun. Participants would then be
instructed to switch to using the equipped handgun and would then walk up to a target
and shoot it. After a while the targets would fall down, and participants could move onto
a new one. In this particular level design, the fallen targets would pop back up after a
certain amount of time, so participants could “cycle” through the targets for their twenty-
minute sessions. The shooting controls were that of a normal first-person shooter, where
participants could control avatar movement and the game camera with analog sticks on
the controller and could aim down sights and shoot the gun using the triggers on the back
of a controller (used an Xbox One wired controller for the controller conditions).
In Golf Club VR, participants had much of the same controls as they did for the
VR shooting game. However, this time the VR controller acted like a golf club. Within
the head display, participants saw a first-person perspective of a golf club being held in
front of them, and the club would move in response to the participants moving the Vive
controller. Participants could hold the controller much like one would hold a real-life golf
club (again, RAs were instructed not to “coach” participants) and would swing at the golf
ball presented to them in the head-mounted display. For the controller version of the golf
game, participants saw a third person-view of the avatar they were controlling. Rather
than the standard first-person controls seen in the shooting condition, participants
controlled the avatar with the left analog stick on the controller (could have them adjust
their orientation to the left or right) and they controlled the golf club with the left analog
stick. To swing the club, participants would pull back on the analog stick and then push it
16
“through” much like the pull-back and follow-though on a golf swing. The harder and
faster one did this with the analog stick, the harder you would hit the ball in the game
(again, analogous to swinging hard with a golf club in real-life).
Self-Reported Efficacy
The Efficacy of Shooting Behavior was previously developed by Carnegy and
Anderson (2008) and used for this study. Items on this scale are rated on a 1 to 7 Likert
scale (1= strongly disagree, 7 = strongly agree), with high scores showing a positive
attitude about guns, high usage of guns, and high confidence and openness to use guns.
The scale is broken up into 17 questions, some sample questions include,
“holding, aiming, and accurately firing a real handgun is easy to do “if needed, I would
have no problem firing a gun at another person,” “I have the ability to hold, aim, and fire
a real handgun,” and “using a firearm to protect loved ones is acceptable to do if they are
in danger.” These items have been adapted to measure attitudes towards golfing and
putting as well by merely replacing “handgun” with “golf putter” or “golf club.” In this
study, the reliability of this scale was consistent with previous research (α = .86);
additionally, the appropriate golf version of this scale also showed strong internal
consistency (α = .92).
These findings are relatively close to the reliability found in similar measures
such as The Attitudes Toward Guns and Violence Questionnaire (Shapiro, Dorman,
Burkey, Welker, & Clough, 1997), which was developed to help assess youth attitudes
towards guns and their likelihood to own and/or use a gun. As Shaprio, Dorman, Burkey,
Welker, & Clough (1997) found, the scale was successful in finding that those who
17
scored high on the scale had a 1 in 3 chance to own a gun, while those who scored low
had a 1 in 125 chance of being a gun owner.
Control Variables and Suspicion
Basic demographic data about participants was gathered. They reported their
gender, age, political affiliation, religious beliefs, GPA, and general socio-economic
information. Experience shooting guns and playing golf will be gathered using the
following questions: “Have you ever fired a real a gun before?”, “Have you ever played
paintball before?” “Have you ever fired an airsoft, or bb, gun before?”, “Have you ever
played golf/mini-golf before?” (1 = yes, 0 = no); “If ‘yes’ then how many times?” (About
once a week/month/year; 4-point scale, 1 = once a year, 4 = once a week). If participants
answered “no” on these questions, they were then scored at “0” for all exposure
calculations.
Experience with video games was also measured. Participants were asked to
report, “How often do you play video games (weekly)?” (Never, Sometimes (0-3 hours),
Regularly (5-10 hours), It is my main hobby (10+ hours) 1 = Never, 4 = It’s my main
hobby). Experience with games that feature shooting or golfing was assessed as well.
Participants will be asked, “Do you play a lot of games with guns/golf?” (Rarely,
Sometimes, Often). For these analyses, only “How often do you play video games
weekly” was calculated. Participants also completed the Aggression Questionnaire (Buss
& Perry, 1992) as well, but results from this scale were not used for this particular
analysis.
Additionally, surveys were collected after each training session assessing
participants’ comfortability and enjoyment of the tasks, but those scales have not been
18
scored or analyzed as they were not part of the main hypotheses. (Both versions of the
scale can be seen in appendices i and j, as well as the scales collected after the training
sessions.
During the final session, and before being debriefed, participants were asked a
series of questions to assess their suspicion about the study. Participants were first asked
what they think of the study, which they can freely type a response to, and then will be
asked a series of “yes/no” questions such as: “Were you confused by the tasks or
instructions?”, “Do you think the lab gave away any information the experiment was
about?” and “Do you think there might have been more to this study than you have been
told?” If participants selected “yes” to any of these questions, they were prompted with a
text box to explain further. (All surveys can be seen in appendices k-m).
Procedure
Before arrival, participants were randomly assigned to one of the six training
conditions. When participants arrived, they completed an informed consent document
(see appendix 14) and were told the necessary procedures of the study and explained that
it took place over a multi-session and multi-week period.
After participants consented to the study and agreed to continue, they completed a
baseline survey that gathered their demographic information, past video game exposure,
past golf/gun exposure, and the efficacy and aggression scales.
Upon completion of these surveys, participants then engaged in the shooting and
putting tests. RAs were instructed not to give any feedback to participants, and if
prompted about technique or form (“Am I holding this right?”), to deflect and say
19
“Whatever you are comfortable with” to help ensure that coaching is not a confounding
variable in this study.
After doing these questionnaires and demographics surveys, participants
completed their first test of shooting and putting performance, which is critical to note, as
this is important to our data analytic procedures. Even though participants only
participated in one of the training paradigms (i.e., VR shooting practice), they would go
ahead and be tested on both shooting and putting tests. This effectively meant that for
each participant, we had a de facto “control” to compare them to (themselves on the skill
they did not practice).
The RAs scheduled participants for their next appointment and logged it on a
Google Calendar page. They also had access to a Google Sheet page where participants
were tracked by a randomized ID that told RAs what condition participants were in. RAs
filled out this document at the beginning and end of every session so they could keep
track of when the participant last came in, when their next upcoming appointment was,
and what session of the study they were currently on.
Participants then went through their two to three weeks of training in their
randomly assigned training condition. These training sessions had each participant
engage in the said training condition for 20-minute periods, then completing a quick
survey measuring their attitudes about the task, how well they thought they did, and their
attitudes about their assistants. Again, the research assistants with them were not allowed
to offer help other than initial instructions on how to play the games, as coaching could
become a significant confound in the results if feedback were allowed.
20
After these four training sessions, participants returned to the lab for a sixth and
final session where they performed the same tests on shooting and putting as they did
initially and were scored again on the same criteria. They then retook the same
questionnaires to assess if their attitudes and confidence toward those specific tasks had
changed.
Hypotheses
We hypothesized that all participants would show improvement between the two
different skills tested (putting and golfing) but that the most amount of improvement for
each participant would be seen for the skills tested that participants also practiced (i.e.,
participants who practiced shooting would improve more on their shooting score than
their putting score; H1). We expected this to further express itself based on which mode
of practice (i.e., VR, real-world, controller) the participants were assigned to practice that
skill in. We anticipated that those who practiced the skill “in the real world” would show
the most amount of improvement, followed by those who practiced the skill in VR, with
those who practiced using a controller showing the least amount of improvement. While
we anticipated that all three practice modes would show improvement on the second test
from the first test, we hypothesized that the amount of improvement would tier in that
specific order (H2) based upon an interaction of which skill they practiced (shooting or
putting) as well as the mode they practiced (real-life, VR, controller).
We also anticipated that the golf and shooting conditions would see roughly the
same amount of improvement (H3) compared to each other. This would indicate that
these skills picked up in games are not merely shooting mechanics, but as other research
has shown, that video games can improve a variety of motor skills.
