Tuesday, 21 February 2012

Are most people really driven by achievement? An assessment of the MBTI population level data

Motivational frameworks in gamification 
There isn't a whole lot of research into gamification right now. One of the seminal texts is "Gamification by Design" by Gabe Ziccherman and David Cunningham. I had an immediate reaction to one of the core concepts of the book "the motivational framework game designers use to segment and engage consumers"

Bartle's Player Typology
Ziccherman and Cunningham use the old version of Richard Bartle's motivational typology (Bartle has subsequently released a new version with 8 archetypes instead of 4) as the basis for choosing which gamification elements to include in a gamified system. Bartle's typology was based on the playing habits of Multi User Dungeon (MUD) games, an old style of relatively primitive role playing games. A lot of the games were not much more than a computerised version of Dungeons and Dragons

The four Bartle player types
Bartle broke play styles into four types:
1. Achiever: achievers like 'levelling up' and demonstrating their superior status to other people 
2. Explorer: explorers like roaming around and finding interesting things about the game world
3. Socialiser: socialisers primarily play games to hang out with other people
4. Killer (or influencer): killers like to take advantage of other people. In the game world, they prey on new players, killing them for fun. In the real world, they might be sales people or politicians - people who become successful by influencing others.

Are Bartle's types useful in gamification?
Reading about this framework made me wonder: is a motivational framework based off Dungeons and Dragons players really relevant for the rest of the population? After all, gamification aims to engage non-gaming audiences. MUD players definitely fit into the 'hard-core' gamer category rather than the much larger casual gaming audience

To be fair, Ziccherman and Cunningham do acknowledge the limitations of Bartle's types. They suggest using the types as a rubric rather than a hard and fast rule and they agree with Bartle's argument that the types are not mutually exclusive. But they made a statement that intrigued me:

For the average person, the breakdown might look something like this:
80% socializer
50% explorer
40% achiever
20% killer 

What motivates the average person?
Ziccherman and Cunningham don't supply any evidence for their breakdown of the average person. So I went looking elsewhere. I found this interesting post from Bart Stewart who integrates Bartle's types with Keirsey's widely used temperaments, which in turn map to the Myers Briggs personality profiles.

Although there are many criticisms of the Myer Briggs Type Indicator (MBTI), there is a heck-of-a-lot of data available from the millions of people who have taken the test. The MBTI personality profiles may not stay static over time for individuals (unsurprising in my opinion, people's personalities do change after all) but the aggregated data represents a 'thin-slice' of the personality profiles of the average person in a given population.

Motivational drivers in the USA
The Myers & Briggs foundation claim (I couldn't find a peer reviewed article) to have aggregated representative population-level data from all of the MBTI tests taken in the US (view here). I took this data and chucked in the matching Keirsey types and then had a go at mapping Bartle's player types.

Type % pop Keirsey role variant Keirsey Style Bartle type Bartle subtype
ISTJ 14.47% Inspector Certifying Achiever Planner
ISFJ 11.84% Protector Securing Achiever Opportunist
ESFJ 11.84% Provider Supplying Achiever Planner
ESTJ 10.53% Supervisor Enforcing Achiever Planner
ENFP 7.89% Champion Motivating Socialiser Networker
ISFP 6.58% Composer Synthesising Explorer Hacker
ISTP 5.26% Crafter Instrumenting Achiever Opportunist
ESTP 5.26% Promoter Persuading Killer Politician
INFP 5.26% Healer Conciliating Socialiser Friend
ESFP 5.26% Performer Demonstrating Socialiser Networker
INTP 3.95% Architect Designing Explorer Scientist
ENTJ 2.63% Field Marshall Mobilising Killer Politician
ENTP 2.63% Inventor Devising Explorer Hacker
INTJ 2.63% Mastermind Entailing Killer Politician
ENFJ 2.63% Teacher Educating Socialiser Networker
INFJ 1.32% Counselor Guiding Socialiser Friend

Based on this data, we get this breakdown:

Achiever Socialiser Explorer Killer
53.95% 22.37% 13.16% 10.53%

If the MBTI data are valid, this makes it look like most people are achievers not socialisers as Ziccherman and Cunningham claim. Hmmm..

Motivational drivers in Australia
How about Australia? (Since I live there) Apparently the Australian Association for Psychological Type (AusAPT) released some statistics around MBTI frequencies in Australia. I say apparently because the link is now broken. All I have to go off is a copied and pasted version of the data on IntrovertForum.com.

I can tell you're groaning already at the lack of data integrity here, but let's play along anyway:P Here's what we end up with

Achiever Socialiser Explorer Killer
46.10% 22.00% 14.90% 16.95%

Everything is pretty similar except for the Achiever and Killer proportions. Wow, apparently Australians are more cut-throat and manipulative (killer archetype) and less achievement oriented than the US. This might explain the much higher incidence of 'tall poppy syndrome' in Australia vs the US.

These preliminary results are very interesting but I don't have much faith in the underlying datasets. There are big concerns about the validity of MBTI testing (there are no controls for exaggeration and lying). Furthermore, the dichotomous nature of MBTI testing (you get crammed into discrete categories rather than appearing on a spectrum) means that the magnitude of achievement drive doesn't show up in the data. Finally, given that the MBI didn't publish the methods for their population level data, I'm doubtful whether the datasets I found are truly representative of the populations.

Next steps
What I'd like to do now is find some population level data for the Big Five personality trait assessment. (This looks promising)

Why this research is important
As part of my PhD, I am in the process of choosing which gamification elements to use in a ridesharing system that I hope will change the way Australians travel. I want this to appeal to the mass market. This kind of system needs careful design so as not to alienate socialisers who have no interest in achievement focused game elements.

No comments: