Subjectivity predicts adjective ordering preferences in Hebrew, but lexical factors matter too

Figures & data

Figure 1. An example of a trial. Participants saw either A or B. The target order was arbitrarily defined as the order on the right (for model-fitting purposes; see Figure 2 below).

Figure 2. An illustration of the calculation of the predictors of the fitted model. All predictors were Z-transformed prior to model fitting. Each participant saw either A or B.

Figure 3. Odds of choosing the target order (see Figure 1) by delta-subjectivity (A) and by delta-frequency (B) (see Fig. 2). Numbers on the x-axis represent standard deviations from the mean.

Table 1. Estimates of the logistic mixed-effects regression model: Target_response_ selections∼Subjectivity*Frequency* Length + (1 + Subjectivity|Participant) + (1|item).

Predictors	Odds Ratios	CI	P
Subjectivity	3.31	2.62-4.18	<0.001***
Frequency	0.82	0.65-1.03	0.082^†
Length	0.95	0.80-1.14	0.599
Subjectivity:Frequency	1.21	0.96-1.52	0.111
Subjectivity:Length	1.10	0.89-1.37	0.383
Frequency:Length	1.20	0.89-1.62	0.233
Subjectivity: Frequency: Length	0.74	0.56-0.97	0.028*

***p < 0.001; **p < 0.01; *p < 0.05; †marginal

Figure 4. Odds of choosing the target order by delta subjectivity, in varying levels of delta frequency (left, middle, and right panels) and length (different line types). Numbers on the x-axis and in panel headers represent standard deviations from the mean.