Usability Testing Video Games With Biometrics

What are biometrics?

Biometrics is an automated way of recording player’s physiological data as they play a game. Players have sensors or observation equipment trained on them during their play session and this data is then used to guide questioning, as well as being used as a source of quantitative data for analysis in its own right. Some basic examples are heart rate monitoring (where the player is required to wear a heart rate monitor, often around their chest) and eye tracking (where infrared cameras track where the player is looking on a screen).

These physiological measures will change due to psychological changes e.g. a player’s heart rate will increase and their pupils will dilate if they are scared because adrenaline is released into their bloodstream.

‍

What can they show?

Biometric data can highlight points in gameplay where the player is having a reaction to the game; biometrics cannot currently give a great deal of information regarding the exact nature of this reaction: is it joy, boredom, fear etc. However, the fact that a reaction has occurred can then be used as a marker for an area where the researcher will want to ask the player what they were thinking at this point and why.

‍

Benefits:

One of the major benefits of biometrics is that it allows the player to be left alone during gameplay to just play. A common method of play analysis is to ask the player to ‘think aloud’. In this style of testing the player is encouraged to speak up whilst playing to give the observers an insight into what they are thinking, why they are doing what they’re doing and what they feel about the game.

This process has obvious drawbacks as the player is being asked to act in a way that is not normal. Quiet players may have to be prompted by observers to talk, meaning that their concentration can be broken as they switch from the game to analysing what they think and how to vocalise this. Using ‘think aloud’ can also be intrusive if the player is using a headset, whilst communicating verbally with friends as part of a multiplayer game online.

Biometrics allows player data to continuously be collected as they play but in a less intrusive way. They may have to wear some kind of sensor but once gameplay starts as long as this is not uncomfortable or painful then the user is not likely to notice it.. Observers can then watch this data being collected in real time and use any interesting changes as markers for questions they wish to ask.

For example, if there is a large spike in heart rate at one point in the game the observer can show the player a screenshot of this moment after the play session and ask them to remember how they were feeling at that time and what they think made them feel this. This is obviously not a perfect science but the benefit of not intruding on the gameplay can be considered a greater benefit than the cost of relying on player memory.

‍

What can be measured and why would you want to measure it

Heart rate:

Many different forms of heart rate measurement exist: from simple beats per minute recorders to complex Electrocardiography recorders. There are established measures for cardiovascular biometrics meaning that analysis is more scientifically concrete. It can be adversely affected by external factors such as player movement, meaning it is hard to decipher which reactions are due to gameplay and which are not.

There is also a fairly large time delay between stimuli and reaction – it takes time for the heart rate to change. For certain games it can be a very useful tool. If a game is trying to be very exciting then a measure can be taken of how many times and at what points in the game the players appeared to get excited (heart rate increased).

‍

Galvanic Skin Response (GSR):

Measures the electrical conductivity of skin. Reactions such as fear, joy, apathy etc. lead to noticeable changes in the conductivity of skin. It is relatively non-intrusive and easy to interpret – if there is a spike (up or down) on the graph then there is a reaction; if the graph is flat lining then the players measurable state is not changing (which could be good or bad depending on what the game is designed to do at this current point).

The main disadvantage of GSR is that it has a noisy signal, meaning lots of other factors could be affecting it, for example muscle contraction near the sensors. There is also a large variation in individuals GSR values meaning different people have very different baseline signals and responsiveness.

‍

Electroencephalogram (EEG):

Measures brain activity, the resolution depends on the quality and thus expense of the hardware being used. Is fairly obtrusive as it requires the player to wear some sort of head gear containing the sensors, often trailing cables meaning that it is hard to forget you are wearing it. Not possible to use on some people because of various reasons such as large amounts of hair, wearing a turban etc. It gives objective and quantifiable data that is reliable and repeatable.

However, it is often time consuming to set up and analyse. Often difficult to interpret as there are thousands of possible reasons different areas of the brain will ‘light up’. No way of telling if the response was because of the game or because player was day dreaming about an unrelated subject.

‍

Eye tracking:

Uses infrared cameras to track where the player is looking on a screen. It requires no hardware to be worn by the user so is almost unnoticeable to player. Can be used to show observers in real time where on the screen the player is looking. Aggregated data can be used to create heat maps showing the main areas of the screen being looked at over a period of time. Analysis can be very time consuming and open to interpretation. Useful for showing observers what general areas of UI player is looking at and what is being ignored or not being seen.

‍

Why use biometric player analysis?

Games differ from the standard usability testing requirements of websites, apps and general software because of their purely entertainment centric nature. Games are played for fun and an important point to note is that games are generally selling the player a problem they need to solve. This varies greatly from nearly all other forms of software interaction. When analysing a website the user testing is, generally, used to ensure that as many users as possible can complete the task they wish to, as quickly as possible.

With gaming the player should often lose/die etc. They have to figure certain things out for themselves, often through the process of trial and error. Due to this unique feature of gaming the player’s engagement becomes much more important. Not only should the game be challenging but it should not be so challenging that the player gets so frustrated they give up; neither should it be so easy that no challenge is given and thus no work is needed on behalf of the player.

This is a major generalisation and varies drastically from game to game and genre to genre. A Facebook game designed to take 5 minutes of play a day will be much simpler to ‘win’ than a triple-A FPS that has 30 hours of gameplay in it. The general gaming principles, however, apply to all games. Biometric analysis can assist across all genres but certain biometric measurements may suit one area more than another.

‍