Testing The Validity of Videogame Modifications for Psychological Research

Dublin Core

Title

Testing The Validity of Videogame Modifications for Psychological Research

Creator

Matthew McArthur

Description

An experiment consisting of 40 participants sampled from both Lancaster University and the Sefton area investigated the validity of using video game modifications (Mods) to construct more reliable testing methods for video game research. Mods are additions to video game code that allow the manipulation of various otherwise unattainable features. It is typical of research in this area to use two different games when attempting to manipulate one variable, this often results in a failure to compare like for like leaving various extraneous variables unaccounted for. Building upon the work of Böche (2009), this present research uses a program Garry’s Mod to construct identical variations that were relatable to popular current games while attempting to manipulate a single variable, violence. Participants played both an antisocial and prosocial variation of the same game and were asked to fill out a survey documenting their opinions on the game and its traits. It was predicted that while features such as graphics, maps, difficulty and the player’s performance in the game would not differ, the self reported violence participants experienced between variations would significantly differ. Predicted trends were observed in ratings of violence t(39) = -22.06, p < .001 and whether the player character’s actions were perceived as prosocial t(39) = -14.64, p < .001. Contrary to predictions, participant’s reported enjoyment t(39) = -4.55, p< 0.001 and the similarity of the game mission to the current game market t(39) = -6.37, p < 0.001 also differed significantly between variations. The implications are discussed throughout the paper.

Source

Participants were randomly allocated into two groups. The first group (“antisocial first”) played and answered questions concerning the antisocial version before then playing and answering questions concerning the prosocial version. The order was reversed for participants in the second group (“prosocial first”).
Antisocial first. The group that played as soldiers first consisted of 14 males and 6 females. The ages of this sample ranged from 18 to 31 years with a mean of 22.25 and a standard deviation of 2.49.
Prosocial first. A group of 14 males, 5 females and one individual who abstained from recording gender played the medic scenario first. The ages ranged from 19 to 26 years, with a mean of 22.3 standard deviation of 1.87.
Task, Materials and Apparatus
Several specialized items were required to set up and conduct this research. In order to attract participants, an email was sent to the Psychology students of Lancaster University as well as several flyers positioned around the department (See Appendix). These adverts offered a chance to win a £20 Amazon gift card for participation and provided individuals with the e-mail of researchers in order to set up meetings.
Hardware. Data was collected using a Macbook Pro with a 13 inch screen. The laptop had a 2.5 GHz (Gigahertz) dual core processor, 4 GB (Gigabytes) of RAM (Random Access Memory), 500 GB Hard Drive memory space and an Intel HD Graphics 4000 graphics card. As well as the Macbook, an apple MagSafe power adapter, a wired optical mouse, a mouse mat and a USB powered laptop cooling pad were used (See Appendix).
Software. The Macbook Pro operated the Mountain Lion operating system by Apple and several programs were required to run this experiment.
Questionnaires concerning demographic information and perceptions of video games were created using an online research suite, Qualtrics. The questionnaire consisted of three sections (See Appendix). Firstly, participants were asked to input their demographic information such as age and gender, and information relating to their video game playing habits (“If you play video games, please tell us the name of your favourite video game?”, “Please tell us approximately how many hours per week you spend playing your favourite game?”). Some questions had pre-written responses, for example, the question regarding a participant’s gender could be answered as ‘☻ Male’, ‘☻ Female’ or ‘☻ Prefer not to say’. However, some questions were purposefully left open to response (e.g. How many hours a week do you spend playing video games?). The second and third section of the survey asked each individual to give responses to questions about the game they had previously played. These two sections were identical but the order of each item had been randomised in an attempt to reduce demand characteristics.
The questions of most importance to the current study regarded how violent the participant felt the game they played was, as well as whether the actions of the video game character they controlled could be considered as prosocial or antisocial. Each question measured responses using a 7-point Likert scale beginning at 0 (“not at all”) and ending with 6 (“completely”). The first exception to this rule was that the question ‘To what extent was your virtual character behaving in a prosocial or antisocial manner?’ by using a scale beginning at -3 and ending with 3. The second exception to this rule was an item asking participants to read the Pan European Game Information (PEGI) definitions for age ratings of video games while asking them to estimate an age rating for the game they had previously played, the scale for this question was a 6-point Likert scale with the options 3+, 7+, 12+, 16+, 18+.
To both construct and run the video game, the first program required was ‘Steam’ a digital video game client created by the company Valve to sell and store video games. Using Steam, Garry’s Mod was purchased to create and edit virtual environments and video game levels through a user-friendly interface featuring weapons, characters, miscellaneous objects and commands. A second game, Counter-Strike, was downloaded using Steam and provided the virtual environments that participants played in.
Using Garry’s Mod, two versions of a game were created. In each version, there were 20 non-player characters (NPCs) scattered around the map for the player to locate and interact with; both variations had NPCs in the same starting location and a limited area in which they were allowed to move. When playing as a soldier, participants were told they had been sent to an enemy base and were tasked to eliminate all forces inside using the SMG. Whereas the medic scenario began under the pretence that the player character was a doctor sent to a quarantined facility to treat contaminated patients using a medikit. Players were informed to expect hostility during the antisocial variation, but were also made aware that 20 health vials scattered around the map to prevent the death of their virtual character. For the medic scenario, participants were informed that quarantined individuals purposefully avoid contact with their character, meaning they would often have to pursue the patient to administer treatment. Participants received no indication from the game as to whether an NPC had previously been treated; however, there was no penalty for giving the same NPC treatment twice.
Design
The study consists of a 2 level within-subject design whereby the independent variable is the version of the video game being played by: the soldier version or the medic version. Though both game environments have been designed to be identical, when individuals play as a solider they are asked to find and kill 20 enemy targets, whereas in while playing as a medic they are required to find and treat 20 contaminated patients.
Procedure
Having displayed interest in taking part in the study, participants were contacted by researchers to arrange a time and location to meet. Once a quiet area that would support the use of a computer with a Wi-Fi connection was found, participants were seated at a desk 20 cm from the MacBook.
Before proceeding, participants were given a Participant Information Sheet (Appendix), which informed them that their participation was voluntary, that they could withdraw at any time and that their data was both anonymous and confidential. Participants were then provided with and asked to sign a consent form (see Appendix).
Participants were given a 5-minute tutorial on the video game controls and mission aims. Participants then completed the demographic questionnaire, which was presented to them on the laptop. Participants indicated their name, gender, age and video game playing habits (see Appendix). Once the demographic questionnaire had been completed, participants played the first version of the video game. Participants in the soldier-first condition played the antisocial version requiring them to shoot other virtual characters. Participants explored using only the keyboard buttons ‘A’, ‘W’, ‘S’, ‘D’ and the mouse, each individual navigated an unfamiliar map searching for 20 NPCs and interacted with them using the left mouse button. A time limit of ten minutes was given for both play throughs while the researcher monitored the amount of soldiers killed or patients saved. After each variation was completed, participants answered a questionnaire on their experience playing the game, such as how similar was the mission given to them compared to games they usually play or how violent they felt the game was. Both questionnaires asked the same 9 questions, however the order in which these were presented was randomised in an attempt to reduce demand characteristics.
After completing both games and their corresponding questionnaires, participants received a debriefing sheet outlining the expected findings and participants were given the chance to ask any questions they had about the study. Lastly, each individual was thanked for taking part and asked to provide a method of contact so that they might be informed if they won the Amazon gift card.
Every participant agreed to have his or her data statistically analysed. Raw data was collected from Qualtrics and exported into an excel file, observations taken on video game performance (e.g number of patients saved) was added manually by the researchers. The dataset was then exported as a .csv file to be investigated using R Studio.

Publisher

Lancaster University

Format

.csv file

Identifier

McArthur

Contributor

Lizzie Wardman

Language

English

Type

Data

Coverage

LA1 4YF

LUSTRE

Supervisor

Dr Neil McLatchie

Project Level

MSc

Topic

Social Psychology

Sample Size

Method
Participants
Opportunity sampling was used to recruit participants from Lancaster University and the Sefton area. Forty participants (mean age = 22.28, range = 18 – 31, SD = 2.2) took part in the present study. The gender divide saw 28 male participants with an age range of 19 to 26 years of age. The mean age of male participants was 21.93 with a standard deviation of 1.65. The 11 female participants of the group were between 18 and 31 years of age with a mean age of 22.82 and a standard deviation of 2.98. One individual opted to not report their gender for demographic purposes.

Statistical Analysis Type

t-tests

Files

Citation

Matthew McArthur, “Testing The Validity of Videogame Modifications for Psychological Research,” LUSTRE, accessed March 28, 2024, https://www.johnntowse.com/LUSTRE/items/show/46.