Model Fit Comparison
.

Quadratic Model
\(y =a+ bx + cx^{2}\), where y is Positive Affect and x is Social Time.
Table S3a.1. Linear and quadratic effects of social time.
| Â |
 |
Positive Affect |
| Â |
 |
B |
CI |
p |
| (Intercept) |
 |
-0.00 |
-0.15 – 0.15 |
1.00 |
| Linear |
 |
0.22 |
0.08 – 0.37 |
.003 |
| Quadratic |
 |
-0.11 |
-0.25 – 0.04 |
.150 |
| Observations |
 | 174 |
| R2 / adj. R2 |
 | .062 / .051 |
.

Diagnostics
.
Assumptions
.
Table S3a.2. Omnibus diagnostic statistics for the quadratic model.
| |
Value |
p-value |
Decision |
| Global Stat |
5.90 |
0.207 |
Assumptions acceptable. |
| Skewness |
2.94 |
0.086 |
Assumptions acceptable. |
| Kurtosis |
0.50 |
0.478 |
Assumptions acceptable. |
| Link Function |
0.16 |
0.686 |
Assumptions acceptable. |
| Heteroscedasticity |
2.29 |
0.130 |
Assumptions acceptable. |
Notes. Tests are directional test statistic
.

Influential Cases.
.

Table S3a.3. Most influential cases by leverage, studentized residual, and Cook’s distance.
| |
StudRes |
Hat |
CookD |
| 22 |
-3.5 |
0.01 |
0.03 |
| 74 |
1.1 |
0.17 |
0.08 |
| 134 |
2.0 |
0.07 |
0.11 |
.
Replottig: Influential Cases Removed

Moderation Model
\(y =x + x^{2} + m +mx + mx^{2}\), where y is Positive Affect, x is Relatedness, and m is Social Time.
Table S3a.4. Linear and quadratic effects of social time.
| Â |
 |
Positive Affect |
| Â |
 |
B |
CI |
p |
| (Intercept) |
 |
0.04 |
-0.11 – 0.18 |
.599 |
| Moderation |
 |
-0.14 |
-0.27 – -0.00 |
.046 |
| Observations |
 | 174 |
| R2 / adj. R2 |
 | .261 / .239 |
Plotting: All Observations

Diagnostics
.
Assumptions
.
Table S3b.5. Omnibus diagnostic statistics for the moderation model.
| |
Value |
p-value |
Decision |
| Global Stat |
4.6 |
0.33 |
Assumptions acceptable. |
| Skewness |
0.0 |
0.98 |
Assumptions acceptable. |
| Kurtosis |
0.1 |
0.75 |
Assumptions acceptable. |
| Link Function |
0.7 |
0.40 |
Assumptions acceptable. |
| Heteroscedasticity |
3.8 |
0.05 |
Assumptions acceptable. |
.

Figure S3b.11. Omnibus diagnostic statistics for the moderation model.