X4_rawData.RdThis file contains the raw data from prior inference experiment.
X4_rawData
A data frame with the following columns:
Participant's ID
Unique ID of a stimulus
Indicates trial order.
Participant self-reported native language
The feature preference of the listener ('red things', 'clouds', 'striped things', etc.). It is the corresponding property that the participants are rating.
The value to which the slider (1-6) was adjusted by the participant. The corresponding property that they are rating is stored under pref1-6. Can be in range 0-1. The value indicates how much a listener likes objects with a particular property. Small numbers mean less liked, larger numbers mean property is preferred.
A three-digit code referring to the object that the speaker intends to signal by the utterance (target), or to one of the other two objects present in the scene (obj2, obj3). The first digit of the code refers to shape, the second digit to texture, and the third digit to color of the object. 1 Shape: 1 cloud, 2 circle, 3 square 2 Texture: 1 solid, 2 striped, 3 polka-dotted 3 Color: 1 blue, 2 red, 3 green Example: If the target is a solid blue square the code would be: 311.
Speaker’s utterance. A speaker chooses among the features present in the scene. All possible utterances include: 'cloud', 'circle', 'square', 'solid', 'striped', 'polka-dotted', 'blue', 'red', 'green'.
A six-digit code that indicates to which category the item belongs. The categorization takes into account how many objects the utterance could possibly refer to, and what properties the target object shares with other objects in the scene. Each category consists of 3 tuples of 2 digits and refers to one feature, where the first tuple always refers to the uttered feature, other features are ordered from most ambiguous to less ambiguous. The more objects share the feature value of the target object, the more ambiguous the feature is. In the example below feature 1 is color since the utterance is 'blue', feature 2 is pattern since the target object shares its pattern (solid) with two other objects. Feature 3 is shape since the target object shares its shape only with one other object. Feature order determines the order of tuples. The first digit in each tuple would then denote how many objects share the value of the picked object for the corresponding feature. We also reordered the objects so that the picked object would be the first. In the experiment, the target object could take any place in the sequence. Example: solid blue square, solid blue circle, solid green square. Ambiguity class: 213222. Utterance: 'blue' Target object: blue solid square How to read code: First tuple: 21 (color) 2: two objects (the target object + 1 more object) share color. 1: the target object shares color with the second object Second tuple: 32 (pattern) 3: three objects share the pattern 'solid'. 2: the target object shared the pattern with 2nd and 3rd objects Third tuple 22 (shape) 2: two objects (the target object + 1 more object) share shape. 2: the target object shares shape with the third object. The third object is not a possible competitor for being chosen, since the utterance is 'blue', and the object is green. If a tuple starts with '3' or '1' the second digit would simply denote whether the other objects have different values for that feature or the same (See the second tuple above).
Refers to whether the utterance of the speaker is ambiguous or not. Example: In a scenario with a blue solid circle, a blue striped square and a red polka-dotted cloud. The utterance 'blue' is ambiguous.