Skip to main content
. Author manuscript; available in PMC: 2022 May 4.
Published in final edited form as: Emotion. 2021 Nov 15;21(7):1357–1365. doi: 10.1037/emo0001025

Table 3.

The Effects of Trait Gratitude and Trait Optimism on Daily Outcomes

Separate predictors
Simultaneous predictors
Gratitude
Optimism
Gratitude
Optimism
Comparison
Outcome variable b t p rb (f) b t p rb (f) b t p b t p z p
Physiology
 Heart Rate −1.18 7.13 <.001 .13 −1.25 7.51 < .001 .19 −.71 3.52 <.001 −.85 4.23 <.001 .42 .676
 Systolic blood pressure −1.12 3.93 <.001 .05 −1.02 3.53 < .001 .02 −.81 2.37 .018 −.57 1.63 .103 −.40 .689
 Diastolic blood pressure −.94 4.69 <.001 .09 −.58 2.82 .005 .05 −.91 3.75 <.001 −.07 .27 .786 −1.96 .051
Health behaviors
 Sleep quality .09 8.45 <.001 .24 .13 11.66 <.001 .34 .03 2.65 .008 .11 8.34 <.001 −3.27 .001
 Moderate physical exercise .14 4.26 <.001 .12 .15 4.44 <.001 .14 .09 2.16 .031 .10 2.50 .013 −.21 .832
 Vigorous physical exercise .10 3.63 <.001 .08 .09 3.32 <.001 .10 .07 2.16 .031 .05 1.57 .116 .32 .752
Stress
 Stress frequency* .87 6.45 <.001 .80 10.95 <.001 .99 .44 .658 .80 8.83 <.001 4.44 <.001
 Stress intensity −.05 3.73 <.001 .06 −.10 8.45 <.001 .22 .02 1.11 .268 −.11 7.64 <.001 4.98 <.001
Expectations and reflections
 Expectations for day .22 12.55 <.001 .45 .24 13.46 <.001 .49 .13 6.34 <.001 .17 7.88 <.001 −.92 .356
 Best part of day .44 15.68 <.001 .39 .31 10.60 <.001 .26 .39 11.55 <.001 .09 2.63 .009 4.97 <.001
 Worst part of day −.23 6.93 <.001 .20 − .41 12.21 <.001 .38 −.02 .40 .690 −.40 9.92 <.001 5.45 <.001
 Appreciation toward others .65 14.37 <.001 .39 .40 8.32 <.001 .24 .62 11.53 <.001 .06 1.09 .275 5.88 <.001

Note. In the far right columns, we created a contrast that subtracted optimism from gratitude. A positive value indicates that the gratitude coefficient is more positive than the optimism coefficient, and a negative value indicates that the optimism coefficient is more positive than the gratitude coefficient. Effect size estimates were calculated using a method explained by Rights and Sterba (2019). The rb(f) statistic is analogous to the square root of the reduction in variance method initially described by Raudenbush and Bryk (2002), akin to a correlation. Due to space constraints, 95% confidence intervals are reported in supplemental Table 1.

*

For the models that included stress frequency as a dichotomous outcome, we calculated odds ratios using a multilevel model for binomial outcomes (odds ratios and z statistics replace b and t statistics for this outcome).