Students are continually exposed to a variety of stressors during their academic career, and this can have significant negative effects on their mental health and subjective wellbeing. In this paper we explore how gamified persuasive interventions can promote engagement in performing random acts of kindness to improve wellbeing and help students manage stressors more effectively. In a pilot study we investigated how participation levels in a gamified persuasive intervention that promotes random acts of kindness at University, are influenced by (1) different persuasive message types, and (2) different game challenge categories. Furthermore, we analysed the impact on behavioural intention by comparing pre-intention and post-intention to perform random acts of kindness. Participants were assigned 5 different quests each morning, for two days, and asked to complete as many as possible by the end of each day. Participants were divided into 2 groups and received different types of persuasive notifications during the day: Group A received messages that set out group goals and used the social comparison strategy, while Group B received messages that set out individual goals and used the self-monitoring strategy. The findings from the pilot study will inform the design of a larger study to investigate persuasive game-based interventions for subjective wellbeing.
BACKGROUND
The number of students in UK Universities who require mental
health care is increasing very rapidly and the problems they
experience are becoming more and more complex [
Despite the growing interest in persuasive games, there
remains a need for further research into their application in the
wellbeing domain and design of games which promote
happiness. Games may facilitate the integration of preventive
models and enhance current proactive strategies. Using technology
to promote wellbeing would encourage initiative, empower
individuals, promote self-care and improve self-management
skills. Approximately 54% of students who experience mental
ill health, feel nervous about receiving care and do not seek
support from their institution or local practitioners [
Our wider research project investigates the design of a
persuasive game for preventing mental health problems and
improving subjective wellbeing in a student population [
STUDY DESIGN The aim of this pilot study was to investigate how engagement levels are influenced by different game challenge categories and different persuasive notification types. We also wanted to explore how gamified persuasive interventions affect behavioural intention towards performing random acts of kindness. The experiment ran over the course of two days, with participants being provided with 5 daily quests each morning, a persuasive notification during early afternoon and a daily questionnaire in the evening. A pre-questionnaire and post-questionnaire were also completed by the participants. To achieve our aims, all quests involved performing small random acts of kindness and were centered around 5 key categories that inspire positivity and promote wellbeing: being positive, having meaningful interactions, expressing gratitude, being helpful and cheering up or encouraging others.
Participants were divided into 2 groups and received different
types of persuasive notifications during the day: Group A
received messages that set out group goals and used the social
comparison strategy, while Group B received messages that
set out individual goals and used the self-monitoring strategy.
Numerous persuasive strategies have been identified to
influence behavioural determinants, in order to promote behaviour
change. For example, Fogg [
Research Questions The pilot study was designed to investigate the following 3 main research questions: 1. How do different persuasive message types affect participant engagement? 2. How do different game challenge categories influence participant engagement?
Participants For the pilot study, we recruited a total of 10 unique participants to take part in experiment (5 females and 5 males, age ranges between 24 and 43 years old). Participants’ geographical territories of origin were Asia and Pacific (2 participants), Europe (4 participants), Africa South of the Sahara (2 participants) and The Middle East and North Africa (2 participants). Participants were postgraduate research students and postdoctoral researchers, recruited from the Computing Science Department at University of Aberdeen. Participants reported that they generally played phone or computer games a few times per year (5 participants), a few times per week (2 participants), every day (2 participants) and almost never (1 participant). Participants were not offered any monetary payment or reward to take part in this pilot study. Table 1 shows demographics and group division of participants.
Participants took part anonymously and did not know there
were 2 groups or which of the groups they were assigned to.
However, given some participants shared offices and the public
nature of some of the tasks, participants may have become
aware of the identity of other participants. This may have had
some influence, which we tried to minimise by participants not
knowing they have been distributed into groups. Additionally,
some participants knowing the experiment leader, and hence
the source of the persuasive messages, could have had an
impact (see [
Pre-Questionnaire
After the collection of demographic information, participants
were asked to describe their happiness level on a scale from 1
(not very happy) to 7 (very happy) at the time of completion.
The Subjective Happiness Scale [
Being Positive I will try to make a spontaneous nice gesture for someone;
Day 1 An email notification was sent to participants, outlining the 5 daily quests for Day 1. A detailed description is provided in Table 2 (Q1 to Q5). Materials required by participants were made available in the indicated locations. Early afternoon, a reminder was sent to the two groups. Participants in Group A were told that there would be a fruit and cake reward if all those who are taking part in the study cumulatively reach a target of at least 30 quests completed in that day. Participants in Group B were told that there would be a fruit and cake reward if they individually reach a target of at least 4 quests completed in that day. At the end of Day 1 participants filled in a questionnaire describing their experiences. For each quest participants were asked if they have completed it. If the answer was "Yes", participants were asked further questions about the quest, as detailed in Figure 1. If participants answered "No", they were asked to explain why they have not completed those quests. Day 2 Participants were sent a new email notification in the morning, containing the daily quests for Day 2. A detailed description is provided in Table 2 (Q6 to Q10). An encouraging message was included in the notification for each group. Participants in Group A were shown a Top 5 Leaderboard displaying participant codes and numbers of quests completed on the previous day and encouraged to maintain their record or even improve it. For Group B, the message informed participants of their own performance only on the previous day and encouraged them maintain or improve this during the course of Day 2. In the evening, participants filled in the questionnaire which asked them which of the tasks they have completed, what motivated them to take part, which tasks they did not do and why. Post-Questionnaire In the post-questionnaire, participants were asked to describe how the different reminders and encouragement messages influenced them when completing the quests during the two days. We reassessed the attitude towards performing random acts of kindness using the same semantic differential scale as in the pre-questionnaire. Finally, we measured participants’ post behavioural intention to perform the 8 acts of kindness described in Table 3 in the upcoming 2 weeks.
RESULTS Having described our approach, we now present our findings. The results are structured around the research questions we investigated in the pilot.
Q1: Influence of persuasive message types A total of 75 challenges have been completed during the pilot study. On Day 1 participants cumulatively completed 41 challenges (20 challenges were completed by Group A and 21 challenges by Group B). On Day 2 a total of 34 challenges have been completed (both Group A and Group B completed 17 challenges). Table 4 summarizes the mean and standard deviation in changes of participation levels for each group over the two days of the pilot. For both groups there is a negative mean change in participation over time potentially motivated by time constraints and other commitments of participants. Figure 2 shows the distribution of completed quests for Group A and Group B based on different persuasive message types. Based on qualitative responses, the persuasive messages influenced both Group A and Group B to complete quests. However, there is a larger cluster of participants from Group B who individually complete more quests than in Group A: the median number of quests completed in Group B is 4 compared to 3 in Group A. This suggests that individual goals and self-monitoring may lead to increased levels of participation, in comparison to group goals and social comparison. However, this result does not reflect how participants perceived the messages influenced them. Overall, participants in Group A reported that the messages had a stronger impact on their performance than those in Group B.
On Day 1, participants in Group A reported that the persuasive notification motivated them to attempt to complete a higher number of quests. Two participants were encouraged by the potential reward offered if the group goal was achieved (PA5: "I wanted to complete more quests because I like cake"; PA4: "I thought fruit and cake can be a good opportunity to share ideas and feelings"). Participants also mentioned that the notifications allowed them to maintain progress (PA2: "I was reminded that there was a task I had not done that I could complete"). One participant (PA1) said that the reward was not a motivating factor as they "enjoy doing good things". Similarly, the majority of participants in Group B (PB2, PB3, PB4) reported that the persuasive message had very little influence on them as they "were not expecting a reward for doing acts of kindness".
On Day 2, all participants in Group A said that the leader board encouraged them to try and maintain or even improve their performance (PA2: "It was motivating to see I had done so well"; PA1: "I wanted to maintain my position in the top"; PA5: "I wanted to do better"). The social comparison strategy influenced one participant who felt motivated to see that other participants are also performing well (PA3: "I can see a lot of good people around me. I am proud of it"). Most participants in Group B reported that the self-monitoring strategy had very little influence on them (PB5: "I would have tried to do my best on the second day too"; PB1: "I was already motivated to complete the quests"). One participant (PB3) said that the notification "brought competitive spirit" to the experience. Q2: Influence of game challenge categories To evaluate each of the 5 key challenge categories we analysed participants’ responses on how happy the challenges made them feel and how enjoyable, motivating and meaningful they found the quests to be. In general, participants said that the quests made them feel happy, were enjoyable and meaningful. However, the quests were considered to have a lower impact on the participants’ motivation. Figure 3 shows how participants engaged in different challenge categories and Figure 4 shows how participants rated the quests in the 5 key categories. Being Positive Quest Category Participants indicated that quests from the "Being Positive" category were the most efficient at providing a feeling of happiness and meaningfulness. The quests in this category were also perceived as enjoyable and motivating, more than most of the other challenge categories. Participants mentioned that they were motivated to complete this type of of challenge mostly due to a pre-existing desire to acknowledge a colleague or member of staff, based on personal experience (PB1: "I wanted to tell a member of staff that they have improved a student’s experience", PB5: "Staff is great and I have things to thank them for", PA4: "I am being thankful for the people in the department"). Respondents were also encouraged by the fact that the quest was to be completed anonymously, considering that this makes the process of recognizing a colleague or member of staff for their merit easier (PB5: "I like to acknowledge and thank staff for the help I get, and it’s even better anonymously"; PB1: "This is a good way to give someone praise without the anxiety of a response"). Participants argued that one issue is that this challenge category requires a certain amount of creative effort to complete (PA1: "I had to think about what to write", PA2: "I intended to say something motivational or inspiring but it was difficult at first"). On each of the days, only 3 out of 10 participants did not complete the challenge belonging to this category, with "time constrains" being the reported cause. Meaningful Interactions Quest Category When asked to introduce themselves to someone new or to ask someone how they were, participants indicated that their engagement levels were mainly influenced by their perception of the meaningfulness of the challenge. Overall, this category of quests was considered the least effective at giving participants a feeling of motivation. It is important to note, however, that on Day 1, this challenge category has the highest percentage of non-completion throughout the entire duration of the experiment, with only half of participants completing it. The main reason for this is that some participants felt it was difficult to engage in conversation with strangers (PB5: "Personally, I find it difficult to strike a conversation with new people"). An interesting observation is made with participants who although reluctant, completed the task (PA1: "Even if I didn’t like introducing myself to a stranger at first, it wasn’t that bad", PB4: "Randomly asking someone how they are might seem strange, although it could turn into a conversation starter"). The overall difficulties that some individuals encounter when communicating with strangers may account for lower perception of benefits that participants gain from this type of quest.
Expressing Gratitude Quest Category Participant responses indicate that the Expressing Gratitude category was very well received by most, based on existing reasons to be grateful towards colleagues. These quests provide the drive for participants to engage themselves in recognizing the efforts of those around them and experience feelings of genuine gratefulness (PA4: "I felt the gratefulness coming with thanking people", PB1: "Feeling genuinely thankful towards some people", PA3: "This may be the most enjoyable quest from my perspective", PB5: "I love thanking people", PA2: "It is one of the best way of appreciating other people."). This type of challenge was also the most engaging, with only three participants failing to complete the quest over the two days period. Time constrains were listed as the main reason. Participants rank this type of quest highly, with small variations between how happy, enjoyable, motivating or meaningful the experience was, as opposed to how other type of challenges were perceived.
Being Helpful Quest Category When asked to provide assistance to someone, participants that completed the quest reported very positively on how happy the task made them feel and also how enjoyable, motivating and meaningful it was. This type of challenge was received well by participants, as they were motivated to complete it because it offers the possibility of direct benefits for themselves (PB3: "Helping the person, I gain knowledge", PB5: "I took a break from writing my report"). However, the style of the quest poses a significant effort, with some participants having difficulties diverting time from work to complete the task or difficulties in engaging with someone that requires assistance. This was the main reason accounting for six non-completed quests, but was also noted by those who completed it (PA2: "I did not speak to anyone who might need help with something", PB2: "I was busy working"). Considering how well this type of task was perceived by participants, but also the extra effort required to complete it, it is suggested that more complex tasks can provide higher emotional rewards.
Cheering Up Others Quest Category The tasks of cheering up others either by leaving a joke or drawing was perceived by participants as the least motivating or meaningful of all the categories. However, participants agree that this category of task was enjoyable and made them feel happy. The task received good overall participation with only five non-completions due to time constrains as the main reason for not completing the quest. Participants considered the challenges to be motivating, as it required them to do an activity that they generally consider enjoyable (PB1: "I enjoy jokes and puns", PA4: "It is very funny, and it may make people smile as it made me"). Some participants noted that this type of challenge can be seen as a strenuous creative task (PA2: "It is difficult to think of something to draw", PA3: "A little bit frustrating. It was hard to make a joke", PB5: "I generally doodle stuff but it doesn’t feel like a meaningful or motivating action"). This characteristic reflects in the low motivation and meaningfulness mentioned by participants. Q3: Impact on Behavioural intention Overall, behavioural post-intention shows an increase from the baseline pre-intention, suggesting that a persuasive gamebased intervention can encourage people to perform more random acts of kindness. A Paired Samples t test was used to compare the behavioural pre-intention and post-intention. There was a significant average difference between pre-intention and post-intention scores (t(9) = 2.484, p < 0.05). On average, post-intention scores were 1.02 points higher than preintention scores (95% CI [.091, 1.948]). Figure 5 shows means of pre-intention and post-intention to perform acts of kindness pertraining to the challenge categories.
CONCLUSIONS AND FUTURE WORK
Our findings in this pilot study help us conclude that
different persuasive message types and different quest categories
influence the level of engagement in game challenges that
promote random acts of kindness. The results provide us with
an indication of how behavioural intention can be improved
by engaging with the persuasive game.
In general, employing individual targets and self-monitoring
as persuasive strategies had a higher practical impact on the
level of engagement than group targets and social comparison.
Overall, participants found all challenge categories enjoyable,
motivating and meaningful, but expressed a preference for
quests belonging to the "Being Positive" category.
A future study could build on the findings from the pilot
experiment and investigate the influence of persuasive message
types and challenge category in a virtually-delivered
persuasive game. A larger number of participants will be recruited
and the influence of other characteristics will be measured
as well (e.g. personality, wellbeing level). For example, it
has been found that adapting messages to personality may
impact persuasiveness [
ACKNOWLEDGMENTS The authors would like to acknowledge and thank all the volunteers who participated in this pilot study and provided helpful comments. The first author is funded by an EPSRC doctoral training grant.