2
\$\begingroup\$

I made a game where I want to check how players act depending on the type of used minimap (minimal vs. path to next task is drawn etc). I came across the Bartle taxonomy of player types for this purpose.

I was able to find the questions data thanks to this question: Bartle taxonomy formula

But I struggle now on how to run this test with people. In the data set are 39 questions. @Trevor Powell mentions in his answer that this means an uneven split between the types:

In the original question set (which is used verbatim in both of the links posted in the question), 'Socialiser' and 'Killer' are each available 20 times, while 'Achiever' and 'Explorer' are each only available 19 times. This biases test results slightly towards producing higher 'Socialiser' and 'Killer' values.

The online questionaires I find use only 30 questions like:

http://matthewbarr.co.uk/bartle/

http://matthewbarr.co.uk/bartle/index.php

I noticed that they have different questions on every run. Like it says on the Wikipedia page, they are randomly selected, and this matches what's on http://www.andreasen.org/bartle/ (you get redirected after 10 seconds).

If I want to get a player type for my participants, it seems odd to me that everyone gets different questions. My plan would be to select 30 questions and ask every participant the same questions.

What is the correct way to administer this survey and determine a player's Bartle type?

\$\endgroup\$
3
  • 1
    \$\begingroup\$ Myself, I would not recommend using Bartle's taxonomy. It's an older model designed for MUDs, and splits players along somewhat artificial bounds. Some more modern options you could try are the Quantic Foundry Gamer Motivation Profile (excellent, but proprietary) or the Five Domains of Play (built on the Big-5/OCEAN model from psychology so question sets are freely available) \$\endgroup\$
    – DMGregory
    Feb 16, 2022 at 11:25
  • \$\begingroup\$ What is your goal? Are you trying to replicate Bartle's results? For that I suppose you should follow the original methodology. If you want to improve upon it, there is no "correct" way we can give you (I guess removing biases would be a goal, also look at alternative models). Are you going to making a survey for the players of your game to help you take design decisions for your game? Then you want questions tailored to your game, and the decisions you need to make. This is a task towards which game designers, psychologists and a social communicators can contribute. \$\endgroup\$
    – Theraot
    Feb 16, 2022 at 12:10
  • \$\begingroup\$ @Theraot I want to compare two types of minimaps. The hypothesis is that player with a high explorer score prefer a minimal minimap while achiever tend to a minimap with high assistant (e.g. path to the next task) \$\endgroup\$
    – Rolfant90
    Feb 16, 2022 at 12:36

1 Answer 1

1
\$\begingroup\$

Social science is no exact science. So trying to implement a Bartle taxonomy test correctly is likely an endeavor which is as futile as it is pointless. You might want to focus on instead is to implementing an own variant of the test in a way which gives you results that are useful for your particular use-case.

So I would advise you to mostly ignore any question catalogs you find on the Internet and instead come up with your own which uses examples which are closer related to the design pillars, mechanics, hooks and themes of the game you developed or plan to develop. That will give you results which are far more useful to you, because they are closer to what kind of role a player would like to play in your game.

\$\endgroup\$
6
  • \$\begingroup\$ I wanted to use the bartletest to get some more information about my players. And then also ask the more specifig questions that are tailored to my research question. I will update my question. I want to compare two types of minimaps. The hypothesis is that player with a high explorer score prefer a minimal minimap while achiever tend to a minimap with high assistant (e.g. path to the next task) \$\endgroup\$
    – Rolfant90
    Feb 16, 2022 at 12:39
  • \$\begingroup\$ @Rolfant90 Then you would ask them questions about how much they enjoy particular exploration, achievement, social and combat aspects of your game. You might also want to use your metrics to cross-reference that data with how much time the players actually spend on those activities in the game. \$\endgroup\$
    – Philipp
    Feb 16, 2022 at 12:42
  • \$\begingroup\$ I would focus on explorer/achiever tendencies, sinde my game has no "killer/socializer" elements (single player). One of of hypothesis is that with a minimal minimap a player will explorer the game more than with a minimap with strong assistance. I think it is also interesting to check if the results for achievers and explorer diverge like that explorer prefer the minimal minimap und achiever prefer the extended minimap \$\endgroup\$
    – Rolfant90
    Feb 16, 2022 at 12:51
  • 2
    \$\begingroup\$ @Rolfant90 When you want to focus on explorers and achievers, then that's one more reason to design your own test. But remember that it is very well possible for a player to enjoy both aspects of your game equally. So having a lot "would you rather achieve or explore" questions might not give you very strong results. \$\endgroup\$
    – Philipp
    Feb 16, 2022 at 12:52
  • 1
    \$\begingroup\$ @Rolfant90 But don't underestimate the socializer appeal of singleplayer games. There are lots of single player games which are enjoyed a lot by socializers. Usually that socializing does not happen in the game itself but rather on the wikis, forums, twitch channels, subreddits, discords and other social media circles around it. But you can still design your game in ways which enable and encourage people to share things about your game online. This does not just improve the appeal to the socializers but also uses their natural tendency to communicate with people as a marketing instrument. \$\endgroup\$
    – Philipp
    Feb 16, 2022 at 12:59

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .