Evaluating the Dimensionality of the Relative Autonomy Continuum in Us and Russian Samples

We conducted a psychometric evaluation of the "relative autonomy continuum" postulated by Self-determination theory (SDT), a continuum whose validity has recently been questioned. We started by a) examining all of the RAI items we could find, across multiple published and unpublished scales; b) extracting the core repeating words and concepts via paired-item paraphrase analysis; and c) expressing all of the resulting concepts in 38 simple, clear new items. We administered the 38 items to multiple Russian and American samples, asking participants to rate their academic motivations. Initial psychometric analyses eliminated several items, leaving 35 items for analysis. The traditional RAI dimensions of amotivated, external, introjected, identified, and intrinsic were confirmed via confirmatory factor analyses, simplex congruency analyses, and multidimensional scaling (MDS) analyses. We also tested a sixth dimension first proposed by Assor, Vansteenkiste et al. (2009), positive introjection, and confirmed its location between negative introjection and identification on the relative autonomy continuum. In addition to confirming the predicted sequence of the items and the six subscales along a primary dimension, MDS analyses also identified a second dimension corresponding to the distance of the item from the center of the continuum, suggesting that using weighting procedures when constructing aggregate motivation scores may be justified. In an attempt to provide the field with a standardized relative autonomy index (SRAI) with known properties, that can be flexibly applied to assess motivation in any and every behavioral domain, we empirically compared several methods of scoring and analyzing the data, focusing on maximizing the associations between academic motivation and subjective well-being. These scoring methods included computing and analyzing each of the six subscales separately; computing and analyzing autonomous and controlled motivation separately; computing a relative autonomy score (autonomous minus controlled motivation); and computing relative autonomy scores in which greater weight is given to subscales nearer to the two extremes of the continuum. Weighted and Unweighted RAI scores predicted SWB equally, indicating that unweighted scoring, which minimizes the number of assumptions made, should be preferred. The positive effect of autonomous motivation was stronger than the negative effect of controlled motivation; intrinsic and introjected motivation were the strongest stand-alone predictors among the 6 sub-scales.