Experimental Design
Methodology
Before outlining the data collection, we describe the two distinct methodologies compared in this study. In the original IOS scale subjects are provided with seven distinct pairs of circles which overlap to a different degree. The overlap represents the strength of relationship between the subject and an individual in question. After considering all seven pairs of circles, subjects then select the pair that most closely represents their relationship with the other individual. A major upside to this methodology is its ease to understand and administer, however, due to the discrete set-up it limits subject’s ability to answer with more nuance. To add the missing nuance researchers have developed the ‘Oneness’ scale. For this measurement the IOS elicitation is combined with the We-Scale, in which subjects answer ‘… to what extent you would use the term “WE” to characterize you and [the individual in question] ’. Even though this combined measurement has been shown to be highly correlated to other measures of relationship closeness and predictive of team production (Gächter et al, 2019), it lacks the simplicity of the original IOS scale.
Thus, when applying the IOS methodology validated by Gächter et al. (2015) researchers frequently solely employ the IOS scale and forego to also elicit Oneness, thereby, maintaining simplicity but losing precision in their measurement or relationship closeness. To address this, we develop an extended version of the IOS scale that combines the simplicity of implementation with more nuance in the measurement. It follows the same graphical style of the original IOS scale but subjects are asked to move a slider in order to adjust the overlap of pairs of circles which appear in a single screen. In order to reduce noise from the choice, the position of the slider is still discrete with now 11 distinct degrees of overlap. Identically to the original IOS, this tool is easy to administer in computerised experiments and provides the subjects with additional opportunities to report relationship closeness in a more finely grained manner.
Data collection
For the data collection we closely follow the approach taken by Gaechter, Starmer, and Tufano (2005). We ask subjects to answer five distinct scales (IOS (original/extended), We-Scale, RCI, LLS, PAM) intended to measure relationship closeness, one of which is either the original IOS scale or our proposed extended version. We then explore in how far ‘Oneness’ or our proposed measure correlate to the different scales, we thereby replicate Gaechter, Starmer and Tufano (2015) whilst in addition exploring the validity of our new methodology.
Moreover, to examine whether the tools are valid across a variety of relationships we also randomly assign subjects to think of a person of one of three relationship levels: close relationship, friend, or acquaintance. Lastly, we also elicit baseline IOS for a stranger, personality traits and self-reported social preferences as additional controls.
Reference
Gächter, S., Starmer, C., & Tufano, F. (2019). The surprising capacity of the company you keep: revealing group cohesion as a powerful factor of team production (No. 2019-16). CeDEx Discussion Paper Series.
Gächter, S., Starmer, C., & Tufano, F. (2015). Measuring the closeness of relationships: a comprehensive evaluation of the 'Inclusion of the Other in the Self' scale. PloS one, 10(6): e0129478.