1

REWARD, REGULATION, AND RISK TAKING ACROSS CULTURESS

Supplemental Materials

Interaction of Reward Seeking and Self-Regulation in the Prediction of Risk Taking: A Cross-National Test of the Dual Systems Model

by N. Duell et al., 2016, Developmental Psychology

Supporting Information

Measurement Invariance

Analysis of measurement invariance across 11 countries for the three self-report measures used in the present study (see S1) was conducted using the alignment method in Mplus V.7.31 (Asparouhov & Muthén, 2014). The alignment method estimates factor mean and variance parameters in each group by incorporating a simplicity function similar to that of the rotation criteria used with exploratory factor analysis (EFA). An iterative procedure determines the largest “set” of groups that contains no significant difference on a given parameter. First, pairwise tests establish the largest set of groups that do not differ significantly ( = .01 to adjust for multiple comparisons). Any two groups that do not differ are “connected.” Second, the average value of the parameter from this invariant group is compared to the value of the parameter of each individual group (whether that groups is part of the invariant set or not). If that comparison suggests non-significant differences (p > .001), the group is added (or kept in) to the invariant set. If that comparison suggests significant differences, the group is removed from (or kept out of) the invariant set. This procedure is repeated until no groups are added or removed for each parameter. Groups that are excluded from the invariant set, therefore, differ from the average value of the invariant group for a given parameter (Muthén & Asparouhov, 2014). Key results are the list of countries with non-invariance for the factor loadings of a particular item (Tables S2 and S3). Additional descriptive information include the fit function contributions, which indicate the relative contribution of a given item to the overall non-invariance across countries and the R2 values which indicate the proportion of non-invariance of a particular item that is explained by variation in the factor means and variances across countries (Table S4).

Summary of Results

Alignment results are presented in Tables S2-S4. All models were estimated using maximum-likelihood estimation. In the models for risk taking and planning, the country with the smallest factor mean was set as a reference group, with the factor mean and variance fixed to 0 and 1, respectively. For sensation seeking, the model was estimated with all factor means and variances free. Such decisions were made based on model identification results in Mplus in order to yield the most accurate, unbiased statistics (please see Asparouhov & Muthén, 2014 for more detail). Based on results for factor loadings reported in Tables S3 and S4, we conclude minimal measurement non-invariance across countries for our self-report items.

1

REWARD, REGULATION, AND RISK TAKING ACROSS CULTURESS

S1. Self-Reported Scale Items

Risk Preference: “How would you compare the good (or pleasures) of the following activities with the bad?”
1. Drinking alcohol
2. Vandalism
3. Riding in a car with a drunk driver
4. Smoking cigarettes
5. Shoplifting
6. Getting into a fight
7. Going into a dangerous neighborhood
8. Threatening or injuring someone with a weapon
9. Having unprotected sexa
Sensation Seeking
1. I like to have new and exciting experiences and feelings even if they are a little frightening.
2. I like doing things just for the thrill of it.
3. I sometimes like to do things that are a little frightening.
4. I’ll try anything once.
5. I sometimes do “crazy” things just for fun.
6. I like wild and “crazy” parties.
Planning
1. I tend to begin a new job without much advance planning on how I will do it. (Reversed)
2. I usually think about what I’m going to do before doing it.
3. I often act without thinking. (Reversed)
4. I hardly ever spend much time on the details of planning ahead (Reversed)
5. Before I begin a complicated job, I make careful plans about how I would complete it.
6. I usually act before I think about what I want to do. (Reversed).

Note.Response options for Risk Preference: (1) Much more good than bad (2) Some more good than bad (3) Some more bad than good (4) Much more bad than good. Responses for each item were reverse-scored so that higher scores represent greater risk preference. Response options for Sensation Seeking and Planning: (1) True (0) False; higher scores indicate greater sensation seeking and planning, respectively. a Item omitted for all participants under 13 years and not included in the questionnaire for Kenya, Jordan, and India.

1

REWARD, REGULATION, AND RISK TAKING ACROSS CULTURESS

S2. Approximate Measurement (Non-) Invariance for Intercepts and Loadings over 11 Countries for Self-Reported Risk Preference Items

Intercepts / Loadings
(1) / (Ch) / It / (Ke) / Ph / (Th) / (Sw) / US / Co / (Jo) / (In) / Cy / (1) / Ch / It / Ke / Ph / Th / Sw / US / Co / Jo / In / Cy
(2) / (Ch) / It / (Ke) / (Ph) / Th / Sw / US / (Co) / Jo / In / Cy / (2) / (Ch) / It / Ke / Ph / Th / Sw / US / Co / Jo / In / Cy
(3) / Ch / It / Ke / Ph / Th / Sw / US / Co / (Jo) / In / Cy / (3) / Ch / (It) / Ke / Ph / (Th) / Sw / US / Co / Jo / In / Cy
(4) / Ch / It / (Ke) / (Ph) / (Th) / Sw / (US) / Co / Jo / (In) / Cy / (4) / Ch / It / Ke / Ph / Th / Sw / US / Co / (Jo) / In / Cy
(5) / Ch / It / Ke / Ph / Th / Sw / US / Co / Jo / In / Cy / (5) / Ch / It / Ke / Ph / Th / Sw / US / Co / (Jo) / In / Cy
(6) / (Ch) / It / Ke / Ph / Th / Sw / US / Co / Jo / In / Cy / (6) / Ch / It / Ke / Ph / Th / Sw / US / Co / Jo / In / Cy
(7) / Ch / It / (Ke) / Ph / Th / (Sw) / US / (Co) / Jo / In / Cy / (7) / Ch / It / Ke / Ph / (Th) / Sw / US / Co / Jo / In / Cy
(8) / (Ch) / It / Ke / Ph / Th / (Sw) / (US) / Co / Jo / In / (Cy) / (8) / Ch / It / Ke / Ph / Th / Sw / US / Co / Jo / In / Cy
(9)a / Ch / It / Ph / (Th) / (Sw) / US / Co / (Cy) / (9) a / Ch / It / Ph / Th / Sw / US / Co / Cy

Note. Numbers represent scale items (see S1). Measurement invariance estimated using alignment method, maximum-likelihood estimation with parameter estimates fixed. Bolded countries inside parenthesis indicate measurement invariance not established in that country for a given item. Ch = China; It = Italy; Ke = Kenya; Ph = Philippines; Th = Thailand; Sw = Sweden; US = United States; Co = Colombia; Jo = Jordan; In = India; Cy = Cyprus. a This item was not asked in Kenya, Jordan, and India. A separate analysis (using all Benthin items and all countries excluding Kenya, Jordan, and India) was conducted to produce results. Measurement invariance results for all other Benthin items were comparable across analyses.

1

REWARD, REGULATION, AND RISK TAKING ACROSS CULTURESS

S3. Approximate Measurement (Non-) Invariance for Intercepts and Loadings over 11 Countries for Self-Reported Sensation Seeking and Planning Items

Intercepts
Sensation Seeking / Planning
(1) / Ch It Ke Ph Th Sw US Co (Jo) In Cy / (1) / Ch It Ke Ph Th Sw US Co Jo In Cy
(2) / Ch It (Ke) Ph (Th) Sw US Co Jo In Cy / (2) / (Ch) It Ke Ph (Th) Sw US Co Jo In Cy
(3) / (Ch) It Ke Ph Th Sw US Co Jo In Cy / (3) / Ch It Ke Ph Th Sw US Co Jo In Cy
(4) / Ch It Ke Ph Th (Sw) US (Co) Jo In Cy / (4) / Ch It Ke Ph Th Sw US Co Jo In Cy
(5) / Ch It (Ke) Ph Th Sw US Co Jo In Cy / (5) / Ch It Ke (Ph) (Th) (Sw) US Co (Jo) In Cy
(6) / (Ch) (It) Ke Ph Th Sw US Co Jo In Cy / (6) / Ch It Ke Ph Th (Sw) US Co (Jo) In (Cy)
Loadings
Sensation Seeking / Planning
(1) / Ch It Ke Ph Th Sw US Co Jo In Cy / (1) / Ch It Ke Ph Th Sw US Co Jo In Cy
(2) / Ch It Ke Ph Th Sw US Co Jo In Cy / (2) / Ch It Ke Ph Th Sw US Co Jo In Cy
(3) / Ch It Ke Ph Th Sw US Co Jo In Cy / (3) / Ch It Ke Ph Th Sw US Co Jo In Cy
(4) / Ch It Ke Ph Th Sw US Co Jo In Cy / (4) / Ch It Ke Ph Th Sw US Co Jo In Cy
(5) / Ch (It) Ke Ph Th Sw US Co Jo In Cy / (5) / Ch It Ke Ph Th Sw US Co Jo In Cy
(6) / Ch It Ke Ph Th Sw US Co Jo In Cy / (6) / Ch (It) Ke (Ph) Th Sw (US) (Co) Jo (In) Cy

Note.Numbers represent scale items (see S1). Measurement invariance estimated using alignment method, maximum-likelihood estimation with parameter estimates free for sensation seeking and fixed for planning. Bolded countries inside parenthesis indicate measurement invariance not established in that country for a given item. Ch = China; It = Italy; Ke = Kenya; Ph = Philippines; Th = Thailand; Sw = Sweden; US = United States; Co = Colombia; Jo = Jordan; In = India; Cy = Cyprus.

1

REWARD, REGULATION, AND RISK TAKING ACROSS CULTURESS

S4. Alignment Fit Statistics for Self-Reported Risk Preference, Sensation Seeking, and Planning

Intercepts / Loadings
Measure & Item / Fit Function / R2 / Fit Function / R2
Risk Preference
(1) Drinking / -49.57 / 0.03 / -23.66 / 0.56
(2) Vandalizing / -29.18 / 0 / -25.45 / 0.50
(3) Car with Drunk Driver / -23.07 / 0.14 / -26.89 / 0.53
(4) Smoking Cigarettes / -32.09 / 0.08 / -23.74 / 0.41
(5) Stealing from Store / -18.77 / 0.93 / -21.87 / 0.58
(6) Getting into Fight / -27.57 / 0.56 / -22.09 / 0.27
(7) Dangerous Area / -31.30 / 0.16 / -24.79 / 0
(8) Threatening Someone / -30.24 / 0.63 / -20.92 / 0.61
(9) Unprotected Sex a / -22.86 / 0 / -10.70 / 0.86
Sensation Seeking
(1) New/Exciting / -30.34 / 0.67 / -27.05 / 0.36
(2) Thrill / -29.74 / 0.59 / -30.79 / 0.49
(3) Frightening / -28.06 / 0.70 / -23.28 / 0.75
(4) Try Anything / -41.52 / 0 / -34.01 / 0.38
(5) “Crazy” Things / -26.97 / 0.58 / -28.82 / 0.26
(6) “Crazy” Parties / -39.09 / 0 / -31.19 / 0
Planning
(1) New Project / -25.81 / 0 / -33.30 / 0
(2) Think Before Doing / -30.06 / 0 / -32.98 / 0.13
(3) Act Without Thinking / -24.68 / 0.80 / -28.56 / 0.55
(4) Planning Ahead / -28.42 / 0.72 / -31.69 / 0.37
(5) Careful Plans / -39.20 / 0 / -26.99 / 0
(6) Act Before Thinking / -33.75 / 0.40 / -45.68 / 0.20

Note. Measurement invariance estimated using alignment method, maximum-likelihood estimation with parameter estimates fixed for risk taking and planning and free for sensation seeking. Fit function values closer to 0 indicate the item contributes less to non-invariance relative to other items. R2 values closer to 1 (0) indicate almost all (non) of the non-invariance of an item can be attributed to variability in factor means and variances across groups. a This item was not asked in Kenya, Jordan, and India. A separate analysis (using all Benthin items and all countries excluding Kenya, Jordan, and India) was conducted to produce results. Measurement invariance results for all other Benthin items were comparable across analyses.

S5. Cronbach Alpha Scale Reliabilities for Self-Report Scales within Individual Countries

Self-Report Scale
Country / Sensation Seeking / Planning / Risk Preference
China / 0.627 / 0.704 / 0.875
Italy / 0.653 / 0.532 / 0.782
Kenya / 0.485 / 0.571 / 0.802
Philippines / 0.568 / 0.578 / 0.812
Thailand / 0.544 / 0.570 / 0.853
Sweden / 0.630 / 0.693 / 0.889
US / 0.642 / 0.613 / 0.836
Colombia / 0.605 / 0.470 / 0.798
Jordan / 0.621 / 0.604 / 0.882
India / 0.775 / 0.731 / 0.822
Cyprus / 0.630 / 0.694 / 0.791

Note.Sensation Seeking and Planning are 6-item scales with response options Yes (1) and No (0); Risk Preference is a 9-item scale with response options on a Likert-type scale ranging from 1—4.

1

REWARD, REGULATION, AND RISK TAKING ACROSS CULTURESS

S6. Means, Standard Deviations, and Score Ranges for all Study Variables within Individual Countries

Risk Preference / Stoplight Risk Index / BART Inflation Ratio (%) / Sensation Seeking
Country / M (SD) / Range / M (SD) / Range / M (SD) / Range / M (SD) / Range
China / 1.36 (.30) / 1—2.67 / 0.46 (.22) / 0—1 / 74.3 (10.39) / 41.35—98.59 / 0.47 (.29) / 0—1
Italy / 1.56 (.36) / 1—2.89 / 0.45 (.17) / 0—0.94 / 69.39 (10.40) / 30.83—93.78 / 0.59 (.29) / 0—1
Kenya / 1.35 (.39) / 1—2.88 / 0.37 (.26) / 0—1 / 67.42 (12.98) / 34.07—97.78 / 0.47 (.25) / 0—1
Philippines / 1.43 (.31) / 1—2.38 / 0.38 (.21) / 0—1 / 70.43 (10.10) / 37.59—93.65 / 0.67 (.25) / 0—1
Thailand / 1.41 (.37) / 1—2.89 / 0.45 (.21) / 0—1 / 69.62 (10.36) / 34.1—98.78 / 0.62 (.25) / 0—1
Sweden / 1.48 (.33) / 1—2.78 / 0.46 (.19) / 0—1 / 76.75 (8.85) / 45.26—96.01 / 0.61 (.26) / 0—1
USA / 1.40 (.36) / 1—2.89 / 0.42 (.19) / 0—1 / 67.26 (12.02) / 33.24—100 / 0.63 (.28) / 0—1
Colombia / 1.43 (.36) / 1—2.78 / 0.43 (.19) / 0—1 / 70.97 (9.80) / 41.03—95.8 / 0.56 (.28) / 0—1
Jordan / 1.22 (.36) / 1—2.63 / 0.25 (.21) / 0—1 / 75.98 (13.80) / 33.67—99.35 / 0.59 (.29) / 0—1
India / 1.42 (.39) / 1—2.88 / 0.53 (.27) / 0—1 / 71.63 (11.93) / 34.63—96.72 / 0.62 (.33) / 0—1
Cyprus / 1.42 (.38) / 1—2.89 / 0.42 (.22) / 0—1 / 69.79 (10.03) / 36.78—93.07 / 0.59 (.28) / 0—1
IGT Reward Sensitivity / Planning / TOL Time to First Move (MS) / Stroop Accuracy
Country / M (SD) / Range / M (SD) / Range / M (SD) / Range / M (SD) / Range
China / 7.94 (20.89) / -60—60 / 0.84 (.23) / 0—1 / 7074.92 (5508.44) / 1142.88—29093.88 / 0.91 (.11) / 0.50—1
Italy / 11.39 (21.57) / -80—70 / 0.65 (.25) / 0—1 / 5413.19 (4603.92) / 1086—29457.13 / 0.93 (.09) / 0.50—1
Kenya / 2.14 (24.91) / -100—100 / 0.79 (.22) / 0—1 / 5219.13 (2336.59) / 1294.88—23546.63 / 0.83 (.16) / 0.42—1
Philippines / 8.32 (20.09) / -50—60 / 0.73 (.24) / 0—1 / 5744.07 (4706.08) / 1255.88—28341.88 / 0.91 (.12) / 0.42—1
Thailand / 4.73 (19.86) / -90—90 / 0.70 (.25) / 0—1 / 3218.64 (2026.08) / 556.75—20650.5 / 0.93 (.10) / 0.42—1
Sweden / 10.60 (21.43) / -100—70 / 0.63 (.29) / 0—1 / 4462.12 (3185.32) / 888.38—28406.25 / 0.93 (.09) / 0.50—1
USA / 9.61 (22.81) / -78.89—90 / 0.74 (.25) / 0—1 / 5919.70 (4603.77) / 595.63—28394.5 / 0.91 (.11) / 0.50—1
Colombia / 4.08 (21.71) / -60—78.89 / 0.67 (.24) / 0—1 / 5244.46 (3437.06) / 1132.63—28304.63 / 0.89 (.12) / 0.42—1
Jordan / -0.95 (22.66) / -100—62.5 / 0.68 (.26) / 0—1 / 4387.55 (3351.62) / 656.13—25898.38 / 0.84 (.16) / 0.42—1
India / 4.30 (20.43) / -100—56.67 / 0.67 (.30) / 0—1 / 5550.03 (4814.22) / 798.88—29338.13 / 0.92 (.10) / 0.42—1
Cyprus / 13.06 (21.44) / -56.67—90 / 0.69 (.28) / 0—1 / 5157.79 (3415.69) / 787.5—24074.88 / 0.91 (.11) / 0.42—1

Note.Self-report items include Risk Preference, Sensation Seeking, and Planning; behavioral items include Stoplight, BART, IGT, TOL, and Stroop. Items used for the risk taking z-score composite include risk preference, Stoplight, and BART; items in the reward seeking z-score composite include sensation seeking and IGT; items in the self-regulation z-score composite include planning, Tower of London, and Stroop

1

REWARD, REGULATION, AND RISK TAKING ACROSS CULTURESS