Researching Power

Challenges to empirically studying power

  • There are multiple kinds of power (e.g., resources, legitimacy) that can sometimes be used to gain other kinds of power, so we can’t operationalize power simply in one way (e.g., authority, outcome control, etc).
  • Power doesn’t necessarily have to be acted on; measuring people’s potential power, and not just power actions they have performed or consequences of those actions, is important.
  • Power changes, so it can’t just be manipulated or measured at one point in time.
  • People do not recognize all uses of power as power, and may change their behavior if it is labeled with reference to power.
  • Experimental control often eliminates choice on participants’ parts, but having choice is necessary to exercising power and to allowing it to have consequences.
  • Power use and need deprivation can have real and dangerous consequences (e.g., death) that a researcher would not want to produce.

The In Game was designed to overcome these challenges

  • Operationalizes several kinds of power at once (e.g., force, resources, legitimacy, obligations, personal attractiveness) using different kinds of game tokens and rules.
  • Potential power can be measured independently from enacted power (behaviors of players) and consequences of use of power (reputations of others, redistribution of potential power, inequality in power distribution, accumulation of potential power, survival).
  • Captures dynamic nature of power because changes are possible throughout the game.
  • Allows researcher to randomly assign conditions but participants are still given choices over how they want to behavior.
  • The choices they make influence the other participants and form the social ecology.
  • Game can simulate consequential and dangerous actions like lack of survival, violence, reproduction, inequality.

 Overview of the In Game

  • Players are given a goal (the most basic one is to “survive” or stay in the game).
  • Players take turns in rounds. On their turns, players have choice about how and whether to act — within constraints of their tokens, the rules, and (in some circumstances) the other player’s situations.
  • Deck of event cards stimulates choices and represents life events.
  • Rules about each kind of game (power) token instantiate different kinds of power (e.g., force, legitimacy).
  • It is not necessary to explain anything about power or to rely on standard symbols and concepts for the game to work.
  • The game permits consequences to self and other players from mutual power use.
  • The game allows researchers to measure potential power (what type of each token all players have), enacted power (behaviors within the game), perceptions of power (players or observers beliefs and attitudes about other players or power in general) and many consequences of collective power use, including subjective well-being and stress, simulated mortality, simulated alliances, fealty, perceived injustice, the development or curtailment of inequality, and emergent cultural norms.
Basic Need Kind of Power Relevant Events Relevant Rules
Wholeness Violence Get another players’ red token Players with 3 more red tokens can take all of other player’s tokens
Consume resources Resources Experimenter’s pool gives out and takes back resources Players with less than minimum # resources are out of game (fail to survive).
Belonging to a community Legitimacy  (These are actions players take that other players like or dislike.) Majority vote can confer or remove legitimacy token from other player
Care from others Obligations  Usually we put at least one event card that states that a player must have another players obligation token People who have given obligation tokens to another player must provide resources to them regularly.
Transcendence Personal attractiveness Share another player’s attractiveness token Risks may accrue to borrowers or lenders

Instantiations of different kinds of power in the In Game:

  • Legitimacy: Peer approval or disapproval; voice
  • Force: Threat of or total disempowerment of another player
  • Resources: Necessary for “Survival” (staying in the game) and frequently exchanged with experimenter’s resource pool
  • Obligations: long-term sharing of resources; calling on others — these tokens are personalized for each player
  • Personal attractiveness as a desirable commodity
  • Arbitrage: A player can only gain access to particular other tokens from a subset of the other players
  • Systematic Disadvantage: Researchers can set the rules or event deck to make certain mishaps more likely for some players than others

The In Game is Flexible

Researchers and teachers are free to adapt the In Game method in ways that suit their purposes. Some suggestions:

  • Omit some types of tokens for simplicity (we have used between 3 and 6 types with college students and other adults).
  • Change particular rules; invent your own new kind of token with rules and events.
  • Make up different kinds of constraints or different kinds of experimental conditions.
  • Provide different goals to players (or to some players).
  • Tell people to play as themselves, on behalf of a group, or to develop a particular kind of reputation with other players.

Contextual Constraints on Fungibility

In our research, we tested whether constraints we imposed by the rules or procedure would change how fungible the kinds of power became:

  • Affiliative constraints — certain players had to exchange obligations only with certain other players, or not  (IG6)
  • Risk for sharing personal attractiveness — players who lent or borrowed a special token had a chance of sitting out a turn for doing so (IG3)
  • Social privileging of agency or communion: Players read that studies show that people like and respect those who are agentic or communal. (IG4)
  • Recommended strategy for play: We told players that those who do best in this game either watch out for themselves or prioritize others. (IG5)

Experimental conditions can be nested

Researchers can establish independent variables that are 1) between game sessions (i.e., all players in the session are in the same condition), and/or 2) within group within session (i.e., more than one person in a session is in a condition, but there are multiple conditions within the session), and/or 3) within players within groups within a session. These can be crossed.

Because players within a session are not independent, data analyses can be done at the session level (e.g., frequency across all players in a session) or should account for non-independence using a hierarchical linear model procedure. In addition, multilevel techniques can test whether, for example, session outcomes like how often “violence” was used in each player’s session influenced how stressed that player felt. In other words, context effects on individuals can be tested.  A different analysis could test whether stress depended on whether that specific player was the perpetrator or victim of “violence.”

Click here for a paper with more details about the In Game

Example Measures (individual, collective)

  • Individual:
    • Individual and dyadic or Social Relations Model: Person-perception and self-perception: trait ratings, trustworthiness, feelings of exploitation, etc.
    • Individual subjective well-being: overall and fluctuations in over turns
    • Accumulation of potential power (tokens at end of game)
    • Power use behaviors (during game, e.g., use of force, alliances)
    • Survival (or length of survival ) per player
    • Perceptions of “culture” of session, fairness of sessions.
  • Collective:
    • Mortality rates
    • Inequality in distributions of power tokens
    • Correspondence between outcomes in sessions (correlations among tokens, symmetry in behaviors)
    • Level of violence in session (e.g., level 2 variables at session: # times violence used)
    • Fungibility: One can measure whether kinds of power are fungible with other kinds of power by testing whether correlations among the tokens are positive. If these correlations are near zero, then one kind of power is not contingent on another and the party this occurs for can be considered less constrained. If these correlations are negative, this implies that the party had to exchange one kind of power for another kind (though not necessarily directly). Although many individuals and groups who are powerful in one way are typically powerful in another, having a deficit in one kind of power could cause a downward spiral in power as well.