Table 1
Methodological details of the 3 studies.
| PRIOR AND GOLLAN (2013) | STASENKO ET AL. (2017) | KLEINMAN AND GOLLAN (2018) | |
|---|---|---|---|
| Participants | 116 bilinguals (4 were excluded)a | 78 Spanish-English bilinguals (2 were excluded)a | 288 Spanish-English bilinguals |
| Number of sessions | Two – a week apart | One | One |
| Task order | Session 1 – language history questionnaire, two similar experimental tasks (language or color-shape) Session 2 – two different experimental tasks (language and color shape)b and even items of MINTc | Language history questionnaire, color-shape switching, language switching, color-word interference test, trail making test, and MINTc | Language switching, language history questionnaire, and the MINTc |
| Experimental tasks and response type | Language switching (digits) – spoken responses Color-shape switching – spoken responses | Language switching (digits) – spoken responses Color-shape – button press | Language switching (picture naming) – spoken responses. |
| Number of trials per condition | 160 trials: 80 single trials (4 blocks) and 80 mixed trials (4 blocks of 20 trials) in sandwich design Switch rate: 50% | 480 trials (half short, 116 ms, and half long, 1016 ms Cue-Target Interval (CTI): 160 single, ~ 160 stay and ~ 160 switch trials. Switch rate: 53% | 324 trials: 216 single trials (2 blocks of 108 trials) and 108 mixed trials (1 block). Switch rate: 33% |
| Reliability analyses | Test retest Internal consistency (even-odd comparisons) | Internal consistency (even-odd comparisons) | Internal consistency (even-odd comparisons) |
[i] a To maximize statistical power we included all participants tested in Prior and Gollan (2013; without excluding 9 Spanish-dominant and 12 Chinese-dominant bilinguals). There were 30 Hebrew-English, 29 Chinese-English bilinguals, and 61 Spanish-English, for a total of 120 participants (four participants were trimmed so that the final sample included 116 participants). In Stasenko et al. (2017), two participants were excluded.
b Half of the participants completed (only) the language task twice in the first session (Training 1 and 2) and once again (Training 3), after completing the color-shape task (transfer task), a week later (hereafter, the language training group). The other half completed (only) the color-shape task twice in the first session and once again, after completing the language task a week later (hereafter, the color-shape training group).
c Multilingual Naming Test (Gollan et al., 2012).
Table 2
Participant characteristics in Prior and Gollan (2013)a.
| HEBREW-ENGLISHb (N = 30) | MANDARIN-ENGLISH (N = 29) | SPANISH-ENGLISH (N = 61) | |
|---|---|---|---|
| Age | 25 | 20 | 20 |
| English self-rated proficiency | 5.8 | 5.9 | 6.5 |
| Other language self-rated proficiency | 7 | 5.4 | 6.0 |
| English MINTc | 24.4 | 28.8 | 29.3 |
| Other language MINT | 31.6 | 25.8 | 23.8 |
| Primary caregiver yrs education | 15.9 | 15.4 | 10.9 |
| Secondary caregiver yrs education | 14.7 | 15.7 | 10.3 |
| Participant yrs education | 13.4 | 13.2 | 13.9 |
| English percentage daily use | 12.4a | 79.9 | 79.6 |
| Age of first exposure to English (yrs) | 8.1 | 5.1 | 4.2 |
[i] a Note that we only describe participant characteristics from Prior and Gollan (2013), and not from the other data sets, because the sample we analyzed herein was substantially different from the original study (i.e., to maximize power in the present study we included all bilinguals including late-learners and those not dominant in the majority language). Language proficiency was rated on a 1 to 7 scale. Ratings presented here are averaged across speaking, listening, reading and writing.
b One participant in this group did not report daily percentage of English use.
c Based on half of the MINT items.
Table 3
Study design (of Prior & Gollan, 2013).
| LANGUAGE SWITCHING | TASK SWITCHING | ||
|---|---|---|---|
| Blocks 1–2 | Single-language blocks (1 English & 1 other, order counterbalanced) | Single-task blocks (1 color & 1 shape, order counterbalanced) | |
| Blocks 3–6 | 4 mixed English/other blocks | 4 mixed color/shape blocks | |
| Blocks 7–8 | Single-language blocks (1 English & 1 other, order reversed from blocks 1 & 2) | Single-task blocks (1 color & 1 shape, order reversed from blocks 1 & 2) | |
| COUNTERBALANCING OF TRAINING AND TRANSFER SEQUENCES | |||
| TIME POINT | TRAINING CONDITION | LANGUAGE TRAINING GROUP | COLOR-SHAPE TRAINING GROUP |
| Day 1 | Training 1 | Language switching | Color-shape switching |
| Training 2 | Language switching | Color-shape switching | |
| Day 2 | Transfer | Color-shape switching | Language switching |
| Training 3 | Language switching | Color-shape switching | |
Table 4
Means and SDs of the different trial types and the switching and mixing costs across tasks in the language and the color-shape tasks.
| PRIOR & GOLLAN, 2013 | STASENKO ET AL., 2017 | KLEIMAN & GOLLAN, 2018 | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| 1ST SESSIONa | TRANSFER TASKb | CTI LONG | CTI SHORT | |||||||
| M | SD | M | SD | M | SD | M | SD | M | SD | |
| LANGUAGE | ||||||||||
| Single | 505 | 64 | 509 | 51 | 520 | 68 | 549 | 75 | 689 | 76 |
| Stay | 554 | 86 | 550 | 75 | 600 | 114 | 675 | 120 | 796 | 107 |
| Switch | 575 | 97 | 577 | 90 | 644 | 127 | 728 | 132 | 841 | 125 |
| Switch cost | 21 | 30 | 28 | 30 | 44 | 38 | 53 | 34 | 45 | 44 |
| Mix cost | 49 | 50 | 40 | 38 | 81 | 68 | 127 | 68 | 106 | 66 |
| COLOR-SHAPE | ||||||||||
| Single | 541 | 69 | 569 | 103 | 526 | 117 | 546 | 122 | ||
| Stay | 601 | 85 | 565 | 214 | 686 | 213 | 920 | 256 | ||
| Switch | 629 | 89 | 679 | 202 | 708 | 217 | 980 | 251 | ||
| Switch cost | 28 | 30 | 25 | 32 | 22 | 51 | 61 | 69 | ||
| Mix cost | 60 | 46 | 73 | 65 | 160 | 147 | 374 | 186 | ||
[i] a First administration of the task on the first day.
b Administration of the task on the second day after training in the other task.
Table 5
Test-retest reliability of single, stay and switch trials and of switching and mixing costs in the language switching task in Prior and Gollan (2013) and internal consistency (correlations between even and odd trials) of the language switching task by study.
| TEST-RETEST | INTERNAL CONSISTENCY | ||||||
|---|---|---|---|---|---|---|---|
| PRIOR AND GOLLAN (2013) | PRIOR AND GOLLAN (2013) | STASENKO ET AL. (2017) | KLEINMAN & GOLLAN (2018) | ||||
| SAME DAY | OVER A WEEK | 1ST SESSIONa | TRANSFER TASKb | CTI LONG | CTI SHORT | ||
| Single | 0.92 | 0.87 | 0.98 | 0.97 | 0.96 | 0.97 | 0.97 |
| Stay | 0.93 | 0.88 | 0.93 | 0.96 | 0.97 | 0.97 | 0.93 |
| Switch | 0.92 | 0.82 | 0.95 | 0.96 | 0.96 | 0.97 | 0.90 |
| switching cost | 0.53 | 0.52 | 0.32 | 0.45 | 0.37 | 0.41 | 0.22 |
| mixing cost | 0.79*c | 0.67 | 0.77* | 0.79* | 0.89* | 0.87* | 0.81* |
[i] a First administration of the task on the first day. b Administration of the task after training on the other task c The only significant difference across domains (i.e., comparing analogous values shown in Tables 5 and 6). * Significantly different from the cell above it (p < .01). # n.s (p > .05).
Table 6
Test-retest reliability of single, stay and switch trials and of switching and mixing costs in the color-shape switching task in Prior and Gollan (2013) and internal consistency (correlations between even and odd trials) of the color-shape switching task by study.
| TEST-RETEST | INTERNAL CONSISTENCY | |||||
|---|---|---|---|---|---|---|
| PRIOR AND GOLLAN (2013) | PRIOR AND GOLLAN (2013) | STASENKO ET AL. (2017) | ||||
| SAME DAY | OVER A WEEK | 1ST SESSIONa | TRANSFER TASKb | CTI LONG | CTI SHORT | |
| single | 0.92 | 0.87 | 0.97 | 0.97 | 0.96 | 0.95 |
| stay | 0.88 | 0.82 | 0.92 | 0.94 | 0.96 | 0.97 |
| switch | 0.90 | 0.76 | 0.91 | 0.95 | 0.96 | 0.97 |
| switching cost | 0.56 | 0.40 | 0.14# | 0.29 | 0.17# | 0.43 |
| mixing cost | 0.53c | 0.51 | 0.70* | 0.82* | 0.91* | 0.91* |
[i] a First administration of the task on the first day. b Administration of the task after training on the other task. c The only significant difference across domains (i.e., comparing analogous values shown in Tables 5 and 6). * significantly different from the cell above it (p < .01). # n.s (p > .05).
Figure 1
Test-retest reliability of language and color shape switching and mixing costs in Prior and Gollan (2011) when tested on the same day and a week apart. Switching costs in the top row, mixing costs in the bottom row.
Figure 2
Internal consistency (i.e., correlating even and odd trials) of language switching and mixing costs across studies.
Figure 3
Internal consistency (i.e., correlating even and odd trials) of color-shape switching and mixing costs across studies (n.b., the axes for mixing costs in Stasenko et al. were adjusted for short and long CTI).
Table 7
ANOVA of the interaction between trial type and parity in language and color-shape tasks across studies.
| F | p | MSE | |
|---|---|---|---|
| Language | |||
| Prior & Gollan (1st task) | <1 | .898 | 353 |
| Stasenko et al. (long) | 1.33 | .267 | 641 |
| Stasenko et al. (short) | 2.98 | .060 | 346 |
| Color-shape | |||
| Prior & Gollan (1st task) | 2.68 | .070 | 508 |
| Stasenko et al. (long) | 1.13 | .325 | 1346 |
| Stasenko et al. (short) | <1 | .859 | 1585 |
Table 8
Summary of main findings.
| SWITCHING COSTS | MIXING COSTS | |
|---|---|---|
| Comparing tasks | Same consistency and reliability across tasks. | Same consistency across tasks. Language more reliable than color-shape when tested twice on the same day and trending in the same direction when tested a week apart. |
| Day effects | No day effect: Similar reliability when tested on the same day and a week apart in both tasks. | No day effect: Similar reliability when tested on the same day and a week apart in both tasks. |
| Comparing mixing to switching costs | Mixing costs were larger and more consistent than switching costs in both tasks. Language task: Mixing costs were more reliable than switching costs when tested on the same day. Color-shape task: Mixing costs were as reliable as switching costs. | |