Grief-Related Rumination

You are a clinical researcher at the Hamilton General Hospital, and your lab is studying how people grieve and cope with the loss of a loved one. Specifically, some people ruminate when they grieve, and so your lab is interested in understanding the different ways in which this grief-related rumination manifests in people. Andrews et al. (2021) recently developed the Bereavement Analytical Rumination Questionnaire (BARQ) to evaluate two dimensions of rumination: 1) the cause of the loss (i.e., root cause analysis – RCA) and 2) how an individual reinvests their time meaningfully following the loss (i.e., reinvestment analysis – RIA).

Your lab is curious about the following questions:

• Do grieving women ruminate more than grieving men?
• Do people ruminate more when the loved one’s death is traumatic?
• Does grief-related rumination vary with the type of relationship with the deceased?
• Does rumination depend on the age of the deceased?
• Does rumination depend on the age of the participant?
• Does rumination depend on the time that has passed since the loved one died?
• Which dimension of the BARQ is more associated with depression?

Similar to Andrews et al. (2021), your lab decides to collect the following information from a questionnaire distributed to 50 respondents:

1. The age of the deceased at the time of death
2. The amount of time that has passed since the time of death
3. The respondent’s current age while completing the questionnaire
4. The respondent’s gender (male, female, or other)
5. The relationship of the deceased to the respondent (i.e., child, parent, spouse, or other)
6. Whether the death was traumatic (yes or no)
7. The average hours of sleep per night after the death (i.e., less than 3 hours, 4-5 hours, 6-8 hours, more than 9 hours)
8. Whether the respondent was prescribed psychiatric medications
9. If the respondent was prescribed psychiatric medications, did that include an antidepressant?
10. Responses to 7 items on the BARQ. Four items form the latent factor RCA, while three items form the latent factor RIA. Respondents rated each of the seven items on a 4-point Likert scale (1 = “Never”, 2 = “Sometimes”, 3 = “Often”, 4 = “All the time”).

To answer the lab’s questions, please run the following analyses.

1. Load the data “P06_dataset.csv”. Run descriptive statistics on the following demographics traits of the sample.
   • What is the mean age of the deceased at the time of death in this sample?
   • What is the mean time that has passed since the time of death in this sample?

• • What proportion of the respondents were female?
  • What proportion of the respondents lost a child?
  • What proportion of deaths were traumatic?
  • What proportion of the respondents were prescribed antidepressant medication

2. Conduct a confirmatory factor analysis (CFA) on the seven items of the BARQ. Items 1-4 should form the latent factor RCA, and items 5-7 should form the latent factor RIA.

• • What is the root mean square error of approximation (RMSEA) of the two-factor CFA?

• • What is the Comparative Fit Index (CFI) and the standardized root mean square residual (srmr) of the two-factor CFA? Use CFI >= .95 and srmr <= .08 as the threshold values.

3. Compare RCA and RIA latent factor means between the following groups. Which comparisons have statistically significant differences in the RCA and RIA means?

• • Respondents who take antidepressant medication vs. those who do not
  • Women vs. men
  • Respondents whose deceased loved one experienced a traumatic death vs. those who did not
  • Respondents who lost a child vs. those who did not

4. The three time variables in the questionnaire may exhibit multilinearity with one another: age of the deceased, current age of the respondent, and the time passed since death. For instance, the age of the deceased and the age of the respondent may be collinear with one another, especially if the relationship of the deceased to respondent is that of a child. To test for multilinearity, assess the variance inflation factor (VIF) of a regression model that includes the three time variables as predictors for RCA and RIA. A VIF of 1 means there is no correlation between predictor variables, while a VIF above 5 indicates a high correlation.

5. Create a scatterplot of the participant’s latent RCA factor score (y-axis) against the age of the deceased child (x-axis).

Files to Download:

1. P06_dataset.csv

References for further reading

Andrews, P. W., Altman, M., Sevcikova, M., & Cacciatore, J. (2021). An evolutionary approach to grief-related rumination: Construction and validation of the Bereavement Analytical Rumination Questionnaire. Evolution and Human Behavior, 42(5), 441-452.