Player Agency in Games: Branches and Choice Feedback

A visible response and a different future are separate things

Player agency in games can raise several questions: what can the player choose, what changes afterward, and how much influence does the player feel? A story can acknowledge a decision immediately while sending every player toward the same next event. Those are separate properties worth examining before counting endings or dialogue options.

For a concrete reading, distinguish the menu of available actions, the subsequent story structure and the player's account of the experience. Evidence about one does not automatically describe the others. A satisfying response can be real feedback even when it does not open a new route through the plot.

The experiment compared three versions of a text adventure

Fendt and colleagues' 2012 study compared a branching text adventure with linear versions offering choice-specific feedback or minimal feedback. There were six binary decisions; two created substantive branches in the branching version. Reported comparisons used 52 branching players on the matched route, 54 linear-feedback players and 44 minimal-feedback players. Participants rated their perceived agency afterward.

This is older research, revisited for its design question. The comparison concerns a short authored story and players' responses to it. It should not be read as an audit of modern role-playing games or a count of how many narrative possibilities a commercial release ought to contain.

The results do not establish that branching is unnecessary

Branching and linear feedback did not differ significantly on four of five story-level agency items at the study's 0.05 threshold. The exception favored branching for perceived alternative outcomes. Linear feedback exceeded minimal feedback on only one of those five items. Branching exceeded minimal feedback on all five.

Failing to detect a difference is not the same as demonstrating equivalence. Nor does this pattern establish that any acknowledgement reliably replaces a consequence. The narrower interpretation is more useful: distinguish the dimensions being rated and preserve the exceptions, rather than turning several questionnaire comparisons into a rule that story structure never matters.

An invented station scene shows the design distinction

Imagine an original story scene in which a courier chooses whether to ask a station attendant for help or inspect a directory. In one version, the next paragraph recognizes that exact action, then sends both routes to the same platform. The response confirms the input, but the courier encounters the same next event.

In a branching version, asking the attendant reveals a delayed train, while the directory leads to an earlier departure from another platform. Now the choice changes subsequent content. This invented example illustrates two structures; it is not another experiment, a tested improvement or a claim that one version will feel better to every reader.

The sample and presentation constrain the conclusion

Recruitment used snowball sampling, the compared groups differed demographically, and the branching comparison selected one route. The authors also identified choose-your-own-adventure framing as a possible influence on expectations.

Those limits matter when interpreting a reported feeling of control. A response after one encounter cannot settle what replaying alternatives would reveal. Likewise, knowing a choice changes a future event and merely expecting it to do so are different kinds of evidence. Avoid treating a questionnaire response as a complete map of the underlying story system.

Ask which part of the choice the evidence addresses

When evaluating a narrative-design claim, identify the choices offered, the immediate acknowledgement, the later changes and the questions players answered. Then ask which of those links the evidence actually supports. A diagram of divergent paths answers a structural question; a player's explanation answers an experiential one.

This distinction supports a precise discussion of choice without promising universal effects or prescribing how every game should be written. The full paper and university record below concern the same 2012 experiment, not independent replications. Their value here is a bounded comparison and a clearer question about what it means for a decision to matter.

Keep exploring

FOLLOW THE SOURCE

Sources & further reading

Have a correction? Tell our editorial team.

✦ NICE. YOU LEARNED SOMETHING.