The Hydra That is a Thesis

In the following blog post I discuss my decision to begin pursuing a joint thesis with Rae Watters and the changes to the way I am approaching my thesis topic. More specifically, my new topic for my meta analysis and the projects I will be working on this semester to conduct my thesis research.

The Endless Abyss of Questions

So another week has gone by and I have been continuing along the journey of attempting to pin down what my thesis will be about. However, to say it's been a simple path would be an understatement. At times it feels like answering a single question just results in several more taking its place. And yet, to say that I am going in circles would be dishonest. Despite the appearance of more and more rabbit holes to explore, progress has indeed been made with every question answered. I just wish I had more time to spend looking down these rabbit holes yet, life waits up for no one. So with that being said, let's get underway with this week's updates. 

Developments In Thesis

Overview

So for this week, I set out with the goal of beginning to work towards my thesis goal of creating a narrative toolkit for designers to use when creating interactive media using dialogue systems that would increase the believability of NPCs. And to be clear, when I say believability, I am referring to Lee and Heeter (2008)’s definition of “the size and nature of the cognitive gap between the character player’s experience and the character they expect”.  This gap when small makes a character be considered believable while when large, likely will interfere with the suspension of disbelief for the player. 

So with regards to my process, four developments can be said:

1. I have now decided to officially partner with Rae Watters, a fellow DAIM Grad Student, to pursue a joint-thesis. For our research, I will be exploring my thesis as described above, while for his research, he will be incorporating those dialogue systems in his narrative design to explore how games can increase player’s acceptance and views towards trans communities.

2. Through discussions of my thesis research with Professor Addy, we have decided to begin exploring how one can measure the believability of an NPC in a games narrative. This would then allow us to determine traits of dialogue systems that increase / decrease believability as well as evaluate my proposed solutions for my thesis.

3. For my meta-analysis, I have shifted to now exploring how suspension of disbelief can increase a story's narrative persuasive outcomes. Through doing so, I will be establishing part of the chain of logic that:

   1. If Suspension of Disbelief increases a narrative’s persuasive outcomes

   2. Then if believability can increase suspension of disbelief or the lack of could decrease suspension of disbelief

   3. Then if my proposed dialogue features can increase / decrease believability

   4. Then my proposed dialogue features can be said to be able to increase or decrease a story’s persuasive outcomes

4. Rae and I have decided for our first prototype we will create a short narrative story that explores using game mechanics as a metaphor for trans stories and inclusion of a gossip system for NPCs to share information with one another.

In the following sections I will break these down in more detail as well as discuss my progress / updates for each course.

Meta Analysis

This week for Meta Analysis, we focused primarily on the problem formation step of a Meta Analysis and learning about the different aspects that go into said step. For example, I learned the difference between a Mediator (one that explains why a variable X impacts Y) and Moderator (one that explains what other variables are impacting Y outside of X) variable. Additionally, we were given examples of past meta-analysises that Dr. Bushman has overseen to serve as inspiration.

After conducting some research into my initial ideas of:

1. Dialogue systems impact on believability

2. Dialogue systems impact on emotional attachment

3. NPC Decision making impact on believability

4. The impact of design decisions on believability

Each one ended up yielding too little (< 10) total studies using a quantitative method to conduct a meta-analysis on the topic. However, after revisiting my definition of believability yielding higher or lower suspension of disbelief (SoD), I considered going the other direction of what can said SoD do? Thus I landed on looking at SoD’s impact on a narrative’s persuasive ability.

Moving forwards into the coming weeks, I will begin starting to gather papers and analyzing them to see if this claim holds true. 

Research Through Design

For this week, I began exploring new approaches to ideation for a project. The first was through the How Might We question approach where a designer asks a series of How Might We questions to better define the project space. For example, with my project space of how do you design a dialogue system that increases believability of an NPC, a potential HMW question could be

How might we create a dialogue system that increases NPCs' social complexity for narrative designers in order to make NPCs seem more believable through better reflecting real life social dynamics.

Additionally, Professor Shen helped walk me through a series of ideation steps of taking a collection of HMW questions and converting them to prototype ideas that could explore the given questions. For example:

Creation of a short story game where the player must identify a killer at a murder mystery without revealing to the killer that the player is seeking them out. However, the player must be careful to not perform an action in front of a party-goer that may go and tip off the killer through the game’s gossip system.

Then from there, identifying what questions those prototypes could answer and what implications such prototypes would have on areas outside of the original problem scope. For example:

Does a gossip system increase a player's perception of an NPC’s social complexity? Does social complexity increase believability? If the prototype does reveal an increase in believability of NPCs due to the added social complexity, what other systems could do likewise? Additionally, what other interactions with virtual avatars could be impacted from such a finding?

After completing my exploration of this topic area, Rae and I decided upon creating a short narrative game with the gossip system as a key feature and will be spending the remainder of this semester iterating on it to improve the perceived social complexity of NPCs due to the gossip system.

Independent Studies

For this week, I met first with Professor Swearingen and then with Professor Addy to further my thesis research.

In my meeting with Professor Swearingen, the two of us discussed the viability of doing a joint thesis with Rae Watters with the central questions being:

• Would we benefit from collaborating with each other?

• How would we collaborate with each other?

• Would we be able to maintain a clear distinction between what we each are hoping to learn through our collaborations?

Based on our conversation, we ended up concluding that due to Rae being a narrative designer primarily focusing on the impact stories can have on players, my research would fit in nicely with his while also helping better inform the creation of the tools I am making. Additionally, due to both of our interests in narrative driven games, it would be easy to focus our collaboration on creation of a game within this genre, especially due to our already extensive background of working on such games together. Finally, due to my research being meant to assist with any narrative designer and aiming to be agnostic to any given story, there would remain a clear distinction between our two thesis topics. Thus, the decision was made to pursue the collaboration.

Additionally, Professor Swearingen and I discussed the importance of learning what is currently being done by dialogue systems. As a result, I was recommended to conduct an additional case study of the different dialogue system mechanics in story driven games. In pursuit of this I created a list of the dialogue systems that I am aware of and will begin attempting to:

1. Identify the system

2. Describe the system

3. Describe examples

4. Reverse-Engineer the system

5. Determine the systems costs and benefits

List of Narrative Systems in Games

Then in my meeting with Professor Addy, we discussed my definition of believability that I have decided upon. More specifically the implications of said definition. Through doing so we ended up with a large list of questions that could then be asked based on such a definition such as:

• How much is believability influenced by real life expectations vs game expectations?

• How does a game’s graphic fidelity influence a player's expectations of an NPC?

• How do allow a developer to make an informed decision about what to include vs not include at the cost of believability?

  
  

From there we began discussing what the next steps would be for my research. In the end, we decided that my research should next focus on the central questions of:

How do you measure believability and determine if it's high or low? What insights can be gathered from using such a measure on current game characters? How can those inform the creation of a framework to intentionally increase / decrease believability?

To answer these questions we decided to perform the following three phases of research:

Phase 1: 

Creation / identification of a method to reliably measure believability

Phase 2: 

Attempt to apply to a collection of games and determine gaps in the methods effectiveness

Phase 3: 

Begin classifying games and their NPCs into different levels of effectiveness of creating believable NPCs 

To begin conducting this research, my goals for next week will be to begin identifying papers measuring believability and identifying the metrics they are using to measure it.

The Wrap Up

Overall, this week has been quite the productive one. Everyday I am reminded why I love being in grad school and having the opportunity to explore these questions. Hopefully through reading these blog posts you have been able to raise your own questions and be inspired to think more upon how you can create systems to increase an NPCs believability.

Until Next Week,

Logging Off

Work Cited

Lee, M. S. and Heeter, C. (2008), ‘Computer science and communication perspectives on character believability in games’, in Annual Meeting of the International Communication Association, Montreal, Canada, 22–26 May.