Correlation Results
The zero-order correlations among the variables are provided in Table 2. As may be seen in the table, ONI scores were positively correlated with all three SD3 scales: Machiavellianism, Narcissism, and Psychopathy. ONI scores were negatively correlated both with RAS scores for participants who were in a committed relationship and with the romance-seeking Likert ratings for participants who were not in a committed relationship. Finally, ONI scores were not correlated with the friendship-seeking Likert ratings.
Table 2 Correlations Among the Study Variables
Variable
|
n
|
1
|
2
|
3
|
4
|
5
|
6
|
7
|
1. ONI
|
788
|
—
|
.42**
|
.11**
|
.51**
|
-.20**
|
-.24*
|
.06
|
2. SD3 Machiavellianism
|
788
|
|
—
|
.24**
|
.55**
|
-.08
|
-.09
|
.16**
|
3. SD3 Narcissism
|
788
|
|
|
—
|
.08**
|
.21**
|
.18
|
.01
|
4. SD3 Psychopathy
|
788
|
|
|
|
—
|
-.38**
|
-.01
|
.21**
|
5. RAS
|
504
|
|
|
|
|
—
|
N/A
|
-.20**
|
6. Romance seeking
|
110
|
|
|
|
|
|
—
|
.61**
|
7. Friendship seeking
|
788
|
|
|
|
|
|
|
—
|
Note. ONI = Orthorexia Nervosa Inventory; SD3 = Short Dark Triad; RAS = Relationship Assessment Scale, only completed by participants in a committed relationship; romance-seeking question only completed by participants who indicated they were single.
*p < .05. **p < .01.
Regression Results with RAS Scores
For the first regression analysis predicting RAS scores, the assumptions of the regression analysis were met. First, the tolerance values (ranging from .71 to .99) were above .2 and the VIF values (ranging from 1.01 to 1.42) did not exceed 10, such that multicollinearity was not present. Second, the partial regression plots revealed that linearity was not violated, and because the predicted x residual scatterplot did not have an obvious pattern, homoscedasticity was not violated either. Third, the residuals remained close to the line on the normal P-P plot, indicating that they were normally distributed. Fourth, the Durbin-Watson value (1.73) was not below 1 or above 3, indicating that residuals were independent. Finally, the Cook’s Distance values (ranging from 0 to 0.19) were below 1, indicating that no influential cases biased the model.
The results of this analysis, presented in Table 3, indicated that RAS scores were negatively associated with ONI scores, such that, as ONI scores increased, relationship satisfaction decreased. However, RAS scores were found to be unrelated to SD3 scores. More importantly, though, the analysis revealed a significant interaction between ONI and SD3 scores in predicting RAS scores. A follow-up simple slopes analysis revealed that the negative association between ONI and RAS scores was significant only at the low and medium SD3 levels (see Table 4). Thus, participants who did not readily relate to examples of Machiavellianism, narcissism, and psychopathy, and exhibited more symptomology of ON, had lower levels of relationship satisfaction. At the high levels of SD3, ONI and RAS scores were not significantly associated.
Table 3 Regression Results with Relationship Satisfaction as the Dependent Variable
Variable
|
B
|
SE
|
LLCI
|
ULCI
|
β
|
t
|
p
|
ONI
|
-.047
|
.014
|
-.076
|
-.019
|
-.169
|
-3.283
|
.001
|
SD3
|
-.023
|
.017
|
-.057
|
.011
|
-.068
|
1.314
|
.189
|
ONI x SD3
|
.002
|
.001
|
.001
|
.004
|
.140
|
3.205
|
.001
|
Note. ONI = Orthorexia Nervosa Inventory; SD3 = Short Dark Triad.
Table 4 Association Between Orthorexia and Relationship Satisfaction at Different Dark Triad Levels
Levels
|
B
|
SE
|
LLCI
|
ULCI
|
β
|
t
|
p
|
Low SD3 levels
|
-.075
|
.023
|
-.120
|
-.031
|
-.241
|
-3.329
|
.001
|
Medium SD3 levels
|
-.082
|
.033
|
-.146
|
-.017
|
-.200
|
-2.509
|
.013
|
High SD3 levels
|
-.013
|
.020
|
-.053
|
.027
|
-.049
|
-0.639
|
.523
|
Note. SD3 = Short Dark Triad.
Regression Results with Romance-Seeking Ratings
For the second regression analysis predicting romance-seeking ratings, the assumptions of the regression analysis were also met. First, the tolerance values (ranging from .93 to .99) were above .2 and the VIF values (ranging from 1.01 to 1.07) did not exceed 10, such that multicollinearity was not present. Second, the partial regression plots revealed that linearity was not violated, and because the predicted x residual scatterplot did not have an obvious pattern, homoscedasticity was not violated either. Third, the residuals remained close to the line on the normal P-P plot, indicating that they were normally distributed. Fourth, the Durbin-Watson value (1.83) was not below 1 or above 3, indicating that residuals were independent. Finally, the Cook’s Distance values (ranging from 0 to 0.24) were below 1, indicating that no influential cases biased the model.
The results of this analysis, presented in Table 5, indicated that romance-seeking ratings were negatively associated with ONI scores, such that, as ONI scores increased, the likelihood of seeking romantic relationship decreased. However, these were unrelated to SD3 scores. In addition, the analysis revealed no significant interaction between ONI and SD3 scores in predicting romance-seeking ratings.
Table 5 Regression Results with Romance-Seeking Ratings as the Dependent Variable
Variable
|
B
|
SE
|
LLCI
|
ULCI
|
β
|
t
|
p
|
ONI
|
-.020
|
.008
|
-.035
|
-.005
|
-.254
|
-2.616
|
.010
|
SD3
|
.006
|
.009
|
-.011
|
.024
|
.070
|
0.724
|
.471
|
ONI x SD3
|
-.001
|
.001
|
-.002
|
.001
|
-.052
|
-0.548
|
.585
|
Note. ONI = Orthorexia Nervosa Inventory; SD3 = Short Dark Triad.