• How did Covid-19 affected people?
• Insights that generalise?
• Advice for future pandemics?
• Collaborations?

• Planned 20-year longitudinal study, currently in its 12\(^{th}\) year.
• Sample frame drawn randomly from NZ Electoral Roll.
• Postal questionnaire (coverage; retention ~ 80%)
• Focus on personality, social attitudes, values, religion, meaning-making, perfectionism, virtues, habits, employment, experiences of discrimination, physical and psychological health, and environmental attitudes …
• Current sample contains 36524 unique people, which is \(>\) 1% of the NZ adult population.
• Here, N = 34782 longitudinal participants who responded to both the 2018/19 and 2019/20 NZAVS waves.

during the past 30 days, how often did…

(a). . . . you feel hopeless; (b). . . . you feel so depressed that nothing could cheer you up; (c). . . . you feel restless or fidgety; (d). . . . you feel that everything was an effort; (e). . . . you feel worthless; (f). . . you feel nervous?

Answers: None of the time; A little of the time; Some of the time; Most of the time; All of the time.”

1. Distress elevation from SARS-CoV-2 health risks (Qiu et al. 2020)(Twenge and Joiner 2020)
2. Distress elevation from economic downturn (Twenge and Joiner 2020);
3. Distress elevation from social isolation (Van Bavel et al. 2020).
4. Distress buffering from greater institutional trust (Sibley et al. 2020).
5. Distress buffering from an elevated sense of civic community (Sibley et al. 2020).

## Marginal Contrasts Analysis
## 
## Level1   |       Level2 | Difference |         95% CI |       SE |  df |     z |      p
## ---------------------------------------------------------------------------------------
## Baseline |   EarlyMarch |      -0.06 | [-0.09, -0.02] |     0.01 | Inf | -5.06 | < .001
## Baseline |       JanFeb |       0.02 | [-0.02,  0.05] |     0.01 | Inf |  1.43 | 0.711 
## Baseline |     Lockdown |      -0.04 | [-0.08,  0.01] |     0.01 | Inf | -2.56 | 0.107 
## Baseline | PostLockdown |      -0.02 | [-0.05,  0.00] | 8.60e-03 | Inf | -2.73 | 0.070 
## Baseline |     PreCOVID |      -0.02 | [-0.03,  0.00] | 3.94e-03 | Inf | -3.88 | 0.001 
## 
## Marginal contrasts estimated at Covid_Timeline
## p-value adjustment method: Holm (1979)

• Manipulate X, measure M and Y
• Regress M on X; Y on X and M

• Y does not affect M
• No 3rd variable on M to Y relationship
• M is measured without error
• Y and M residuals are not correlated (Vuorre 2020)

path_m <- bf(
  mF ~ x + (1 | c | id)
  )
path_y <- bf(
  y ~ x + mF_w + mF_b +
               (1 | c | id)
  )
m1 <- brm(
  path_m + path_y + set_rescor(FALSE),
  data = datF,
  file = here("models/mediation-k6-covid-fatigue")
)

• L4-Lockdown eased distress
• Rest, security, exercise were buffers
• Relevant for future pandemic planning
• contact: joseph.bulbulia@vuw.ac.nz

• Prof Chris G. Sibley & the NZAVS team
• Dr.Matti Vuorre (former VUW graduate, now at Oxford) for his workshop
• NZAVS participants

https://go-bayes.github.io/reports/posts/covid/

library("ggdag")
d1<-dagify(
  k6  ~ covid + commute + work,
  commute ~ work + covid,
  work ~ covid,
  exposure =  "commute",
  outcome =   "k6",
  labels = c(k6 = "Kessler 6", 
             covid = "Covid Timeline", 
             work = "hours work",
             commute = "hours commute"
             ))