**4. Results**

The results are presented here in terms of language preference across with the condition (i.e., participants conversed with another KQSL-ISL bilingual, a KQSL monolingual, and an ISL monolingual). A total of 2754 tokens were included as part of this analysis.

Table 5 shows the individual differences in KQSL use in each condition, and on average (for all three conditions). The use of KQSL lexical signs among bilinguals ranged from 2.8 to 63.6%, with an average of 17.1%. On average, most participants preferred the use of ISL in all conditions, with the exception of one participant—KQSL-ISL bilingual no. 06—who used more KQSL on average (63.6%) than ISL, although less KQSL than ISL in the monolingual ISL condition.


**Table 5.** Individual variation of KQSL.

KQSL lexical sign use varied across participants depending on the condition. In Figure 3, the percentage of KQSL lexical sign use is presented by individual (1–12 on the X axis) and by condition (blue is the bilingual condition, orange is the monolingual KQSL condition, and gray is the monolingual ISL condition). As shown in Figure 3, KQSL was rarely used in the monolingual ISL condition (i.e., all grey bars are low). For the other two conditions, some participants used more KQSL lexical signs in the monolingual KQSL condition than in the bilingual condition, while others showed the reverse pattern. For example, bilingual 01 shows a decline in KQSL lexical sign use as follows: bilingual > monolingual KQSL > monolingual ISL; while bilingual 06 shows a decline in KQSL lexical sign use in a different order: monolingual KQSL > bilingual > monolingual ISL. In some cases, e.g., bilinguals 02 and 10, participants showed a strong increase in KQSL in the monolingual KQSL condition, but overall, showed a preference for ISL. In contrast, several participants did not change their signing across conditions, using ISL predominantly regardless of whom they interacted with (see Table 5).

We conducted a multiple regression analysis to test whether participants' use of KQSL was predictable by condition. Use of KQSL was included as the dependent variable, and condition as the independent variable. Participant was included as a random effect. The results indicated a significant effect of condition on the use of KQSL at a significance level of *p* < 0.001 (1.46 × <sup>10</sup>−40). In the bilingual condition, KQSL lexical signs constituted an average of 24% of the overall lexical sign production, 20.6% in the KQSL monolingual condition, and only 4.8% in the ISL monolingual condition. Table 6 presents the results, including the log odds, number of tokens analyzed, percentage of KQSL lexical signs, and the centered weight (with KQSL lexical signs as the application value). Results with a positive log-odd and a factor weight over 0.5 indicate an increased likelihood that KQSL will be used; while a negative log-odd and a factor weight below 0.5 indicate an increased likelihood that ISL will be used.

**Figure 3.** Percentage of KQSL lexical sign use by individual, presented across conditions.

**Table 6.** Significant Rbrul results (by condition).


Application value: KQSL signs. Significant at *p* < 0.01. 2754 tokens. Random (participant) standard deviation = 1.225.
