Familism_PathsAB_Stigma
Melissa-Ann Lagunas
2023-12-01
# import data
library(readr)
d <- read_csv("/Users/melissalagunas/Desktop/Lab/Contextualizing\ Stigma\ and\ Trauma/LatineData_Subscales.csv")## New names:
## Rows: 259 Columns: 16
## ── Column specification
## ──────────────────────────────────────────────────────── Delimiter: "," dbl
## (16): ...1, SSRPH, SSOSH, MHSAS, PR_PF, PR_A, PR_D, PR_SR, PR_PI, ANG, C...
## ℹ Use `spec()` to retrieve the full column specification for this data. ℹ
## Specify the column types or set `show_col_types = FALSE` to quiet this message.
## • `` -> `...1`# import data
library(readr)
dat <- read_csv("/Users/melissalagunas/Desktop/Lab/Contextualizing\ Stigma\ and\ Trauma/LatineData2.csv")## New names:
## Rows: 259 Columns: 151
## ── Column specification
## ──────────────────────────────────────────────────────── Delimiter: "," chr
## (4): AHELP, GEN_8_TEXT, ETH_7_TEXT, BORN_1 dbl (146): ...1, Progress,
## Duration..in.seconds., Consent, SSRPH_1, SSRPH_2,... lgl (1): EDU_11_TEXT
## ℹ Use `spec()` to retrieve the full column specification for this data. ℹ
## Specify the column types or set `show_col_types = FALSE` to quiet this message.
## • `` -> `...1`# remove NA variables
d <- na.omit(d)#data set with only variables of interest
library(dplyr)##
## Attaching package: 'dplyr'## The following objects are masked from 'package:stats':
##
## filter, lag## The following objects are masked from 'package:base':
##
## intersect, setdiff, setequal, uniond <- data.frame(d$PHFS, d$MHSAS, d$SSOSH, d$SSRPH)
d <- rename(d, familism = d.PHFS, help_seeking_attitudes = d.MHSAS, self_stigma = d.SSOSH, public_stigma = d.SSRPH)#loading packages
library(lavaan)## This is lavaan 0.6-16
## lavaan is FREE software! Please report any bugs.library(semPlot)
library(tidyverse)## ── Attaching core tidyverse packages ──────────────────────── tidyverse 2.0.0 ──
## ✔ forcats 1.0.0 ✔ stringr 1.5.0
## ✔ ggplot2 3.4.4 ✔ tibble 3.2.1
## ✔ lubridate 1.9.3 ✔ tidyr 1.3.0
## ✔ purrr 1.0.2## ── Conflicts ────────────────────────────────────────── tidyverse_conflicts() ──
## ✖ dplyr::filter() masks stats::filter()
## ✖ dplyr::lag() masks stats::lag()
## ℹ Use the conflicted package (<http://conflicted.r-lib.org/>) to force all conflicts to become errorslibrary(psych)##
## Attaching package: 'psych'
##
## The following objects are masked from 'package:ggplot2':
##
## %+%, alpha
##
## The following object is masked from 'package:lavaan':
##
## cor2covlibrary(jtools)
library(interactions)
library(ggplot2)
library(tidySEM)## Loading required package: OpenMx
## OpenMx may run faster if it is compiled to take advantage of multiple cores.
##
## Attaching package: 'OpenMx'
##
## The following object is masked from 'package:psych':
##
## tr
##
## Registered S3 method overwritten by 'tidySEM':
## method from
## predict.MxModel OpenMx
##
## Attaching package: 'tidySEM'
##
## The following object is masked from 'package:jtools':
##
## get_data# Correlation Analysis
table <- apaTables::apa.cor.table(d, table.number = 1, show.sig.stars = TRUE,
landscape = TRUE, filename = "table.doc")## Registered S3 methods overwritten by 'broom':
## method from
## tidy.glht jtools
## tidy.summary.glht jtoolsprint(table)##
##
## Table 1
##
## Means, standard deviations, and correlations with confidence intervals
##
##
## Variable M SD 1 2 3
## 1. familism 3.68 0.99
##
## 2. help_seeking_attitudes 5.54 1.16 .21**
## [.09, .32]
##
## 3. self_stigma 2.32 0.83 -.18** -.55**
## [-.30, -.06] [-.63, -.46]
##
## 4. public_stigma 2.22 0.67 -.12 -.40** .56**
## [-.24, .00] [-.50, -.29] [.46, .63]
##
##
## Note. M and SD are used to represent mean and standard deviation, respectively.
## Values in square brackets indicate the 95% confidence interval.
## The confidence interval is a plausible range of population correlations
## that could have caused the sample correlation (Cumming, 2014).
## * indicates p < .05. ** indicates p < .01.
## psych::pairs.panels(d)
# Checking the table of means, standard deviations, and correlations
table <- apaTables::apa.cor.table(d, table.number = 1, show.sig.stars = TRUE,
landscape = TRUE, filename = "Lewis_Corr.doc")
print(table)##
##
## Table 1
##
## Means, standard deviations, and correlations with confidence intervals
##
##
## Variable M SD 1 2 3
## 1. familism 3.68 0.99
##
## 2. help_seeking_attitudes 5.54 1.16 .21**
## [.09, .32]
##
## 3. self_stigma 2.32 0.83 -.18** -.55**
## [-.30, -.06] [-.63, -.46]
##
## 4. public_stigma 2.22 0.67 -.12 -.40** .56**
## [-.24, .00] [-.50, -.29] [.46, .63]
##
##
## Note. M and SD are used to represent mean and standard deviation, respectively.
## Values in square brackets indicate the 95% confidence interval.
## The confidence interval is a plausible range of population correlations
## that could have caused the sample correlation (Cumming, 2014).
## * indicates p < .05. ** indicates p < .01.
## # quick peak at data
psych::describe(d)## vars n mean sd median trimmed mad min max range
## familism 1 257 3.68 0.99 3.80 3.77 0.89 1 5.0 4.0
## help_seeking_attitudes 2 257 5.54 1.16 5.78 5.66 0.99 1 7.0 6.0
## self_stigma 3 257 2.32 0.83 2.10 2.25 0.74 1 4.9 3.9
## public_stigma 4 257 2.22 0.67 2.20 2.21 0.59 1 4.0 3.0
## skew kurtosis se
## familism -0.76 0.13 0.06
## help_seeking_attitudes -1.13 1.45 0.07
## self_stigma 0.73 -0.05 0.05
## public_stigma 0.19 -0.05 0.04Working the Moderated Mediation
Piecewise assemly of the moderated mediation
Analysis 1: A simple moderation
# Y = self-stigma; X = public stigma; W = familism
Mod_a_path <- lm(self_stigma ~ public_stigma * familism, data = d)
jtools::summ(Mod_a_path, digits = 3)## MODEL INFO:
## Observations: 257
## Dependent Variable: self_stigma
## Type: OLS linear regression
##
## MODEL FIT:
## F(3,253) = 43.036, p = 0.000
## R² = 0.338
## Adj. R² = 0.330
##
## Standard errors: OLS
## --------------------------------------------------------------
## Est. S.E. t val. p
## ---------------------------- -------- ------- -------- -------
## (Intercept) -0.088 0.556 -0.159 0.874
## public_stigma 1.227 0.230 5.343 0.000
## familism 0.243 0.143 1.698 0.091
## public_stigma:familism -0.149 0.060 -2.496 0.013
## --------------------------------------------------------------#graph interaction
interactions::interact_plot(Mod_a_path, pred = public_stigma, modx = familism) +
ylim(1, 10)
#probing the interaction with simple slopes
interactions::sim_slopes(Mod_a_path, pred = public_stigma, modx = familism)## JOHNSON-NEYMAN INTERVAL
##
## When familism is OUTSIDE the interval [6.12, 25.27], the slope of
## public_stigma is p < .05.
##
## Note: The range of observed values of familism is [1.00, 5.00]
##
## SIMPLE SLOPES ANALYSIS
##
## Slope of public_stigma when familism = 2.686265 (- 1 SD):
##
## Est. S.E. t val. p
## ------ ------ -------- ------
## 0.83 0.09 9.41 0.00
##
## Slope of public_stigma when familism = 3.676265 (Mean):
##
## Est. S.E. t val. p
## ------ ------ -------- ------
## 0.68 0.06 10.53 0.00
##
## Slope of public_stigma when familism = 4.666264 (+ 1 SD):
##
## Est. S.E. t val. p
## ------ ------ -------- ------
## 0.53 0.09 6.08 0.00 # The Johnson-Neyman is a floodlight approach and provides an indication of the places in the distribution of W (moderator) that X has an effect on Y that is different than zero. The analysis of simple slopes approach is thought of as a spotlight approach because probes the distribution at specific values (often the M +/- 1SD).In this model that is, overall, statistically significant, we account for about 33% of the variance in self-stigma. Looking at these results we can see that there is a statistically significant main effect of public stigma on self-stigma as well as a statistically significant interaction effect (public stigma:familism [XM])
Analysis 2: Another simple moderation
# Y = help-seeking attitudes; X = self-stigma; W = familism
Mod_b_path <- lm(help_seeking_attitudes ~ self_stigma * familism, data = d)
jtools::summ(Mod_b_path, digits = 3)## MODEL INFO:
## Observations: 257
## Dependent Variable: help_seeking_attitudes
## Type: OLS linear regression
##
## MODEL FIT:
## F(3,253) = 38.379, p = 0.000
## R² = 0.313
## Adj. R² = 0.305
##
## Standard errors: OLS
## ------------------------------------------------------------
## Est. S.E. t val. p
## -------------------------- -------- ------- -------- -------
## (Intercept) 6.965 0.628 11.088 0.000
## self_stigma -0.829 0.237 -3.504 0.001
## familism 0.074 0.161 0.463 0.644
## self_stigma:familism 0.026 0.062 0.418 0.676
## ------------------------------------------------------------#graph interaction
interactions::interact_plot(Mod_b_path, pred = self_stigma, modx = familism) +
ylim(1, 10)
#probing the interaction with simple slopes
interactions::sim_slopes(Mod_b_path, pred = self_stigma, modx = familism)## JOHNSON-NEYMAN INTERVAL
##
## When familism is INSIDE the interval [-3.88, 8.42], the slope of
## self_stigma is p < .05.
##
## Note: The range of observed values of familism is [1.00, 5.00]
##
## SIMPLE SLOPES ANALYSIS
##
## Slope of self_stigma when familism = 2.686265 (- 1 SD):
##
## Est. S.E. t val. p
## ------- ------ -------- ------
## -0.76 0.09 -8.11 0.00
##
## Slope of self_stigma when familism = 3.676265 (Mean):
##
## Est. S.E. t val. p
## ------- ------ -------- ------
## -0.73 0.07 -9.89 0.00
##
## Slope of self_stigma when familism = 4.666264 (+ 1 SD):
##
## Est. S.E. t val. p
## ------- ------ -------- ------
## -0.71 0.10 -7.13 0.00 # The Johnson-Neyman is a floodlight approach and provides an indication of the places in the distribution of W (moderator) that X has an effect on Y that is different than zero. The analysis of simple slopes approach is thought of as a spotlight approach because probes the distribution at specific values (often the M +/- 1SD).Looking at these results we can see that the predictors account for 31.3% of variance in help-seeking attitudes. Only the independent variable (X), self-stigma is a significant predictor. Neither the moderator (familism), nor its interaction with self-stigma (self_stigma:familism [XW]) are signficant.
Analysis 3: A simple mediation
# Y = help-seeking attitudes; X = public_stigma; M = self_stigma
LMedModel <- "
help_seeking_attitudes ~ b*self_stigma + c_p*public_stigma
self_stigma ~a*public_stigma
#intercepts
self_stigma ~ self_stigma.mean*1
help_seeking_attitudes ~ help_seeking_attitudes.mean*1
indirect := a*b
direct := c_p
total_c := c_p + (a*b)
"
set.seed(230925) #required for reproducible results because lavaan introduces randomness into the calculations
LMed_fit <- lavaan::sem(LMedModel, data = d, se = "bootstrap", missing = "fiml")
LMed_Sum <- lavaan::summary(LMed_fit, standardized = T, rsq = T, ci = TRUE)
LMed_ParEsts <- lavaan::parameterEstimates(LMed_fit, boot.ci.type = "bca.simple",
standardized = TRUE)
LMed_Sum## lavaan 0.6.16 ended normally after 1 iteration
##
## Estimator ML
## Optimization method NLMINB
## Number of model parameters 7
##
## Number of observations 257
## Number of missing patterns 1
##
## Model Test User Model:
##
## Test statistic 0.000
## Degrees of freedom 0
##
## Parameter Estimates:
##
## Standard errors Bootstrap
## Number of requested bootstrap draws 1000
## Number of successful bootstrap draws 1000
##
## Regressions:
## Estimate Std.Err z-value P(>|z|) ci.lower
## help_seeking_attitudes ~
## slf_stgm (b) -0.659 0.101 -6.530 0.000 -0.864
## pblc_stg (c_p) -0.236 0.123 -1.921 0.055 -0.484
## self_stigma ~
## pblc_stg (a) 0.694 0.075 9.219 0.000 0.545
## ci.upper Std.lv Std.all
##
## -0.463 -0.659 -0.472
## -0.002 -0.236 -0.135
##
## 0.840 0.694 0.555
##
## Intercepts:
## Estimate Std.Err z-value P(>|z|) ci.lower ci.upper
## .slf_stg (sl_.) 0.780 0.154 5.080 0.000 0.480 1.072
## .hlp_sk_ (h__.) 7.587 0.225 33.737 0.000 7.145 8.021
## Std.lv Std.all
## 0.780 0.938
## 7.587 6.537
##
## Variances:
## Estimate Std.Err z-value P(>|z|) ci.lower ci.upper
## .hlp_skng_tttds 0.927 0.125 7.397 0.000 0.679 1.158
## .self_stigma 0.478 0.042 11.483 0.000 0.396 0.559
## Std.lv Std.all
## 0.927 0.688
## 0.478 0.692
##
## R-Square:
## Estimate
## hlp_skng_tttds 0.312
## self_stigma 0.308
##
## Defined Parameters:
## Estimate Std.Err z-value P(>|z|) ci.lower ci.upper
## indirect -0.457 0.090 -5.069 0.000 -0.648 -0.289
## direct -0.236 0.123 -1.920 0.055 -0.484 -0.002
## total_c -0.693 0.106 -6.528 0.000 -0.888 -0.462
## Std.lv Std.all
## -0.457 -0.262
## -0.236 -0.135
## -0.693 -0.397The a path (public_stigma -> self_stigma) is statistically significant The b path (self_stigma -> help_seeking_attitudes) is statistically significant The total effect (public_stigma -> help_seeking_attitudes) is statistically significant The direct effect (public_stigma -> help_seeking_attitudes when self_stigma is in the model) is marginally significant The indirect effect (ab) is statistically significant The model accounts for 31% of the variance in help-seeking attitudes (DV) and 30% of the variance in self-stigma
Analysis 4: The moderated mediation - a combined analysis
# Specify the mediation model
Combined <- '
# Direct effect (path C)
help_seeking_attitudes ~ c * public_stigma
# Mediation model (paths A and B)
familism ~ a * public_stigma
help_seeking_attitudes ~ b * self_stigma
# Moderation effect on path A
self_stigma ~ d * familism
# Moderation effect on path B
help_seeking_attitudes ~ e * familism
'
# Fit the model
fit <- sem(Combined, data = d)
# Summarize the results
summary(fit)## lavaan 0.6.16 ended normally after 1 iteration
##
## Estimator ML
## Optimization method NLMINB
## Number of model parameters 8
##
## Number of observations 257
##
## Model Test User Model:
##
## Test statistic 91.302
## Degrees of freedom 1
## P-value (Chi-square) 0.000
##
## Parameter Estimates:
##
## Standard errors Standard
## Information Expected
## Information saturated (h1) model Structured
##
## Regressions:
## Estimate Std.Err z-value P(>|z|)
## help_seeking_attitudes ~
## publc_stgm (c) -0.231 0.090 -2.563 0.010
## familism ~
## publc_stgm (a) -0.176 0.092 -1.908 0.056
## help_seeking_attitudes ~
## self_stigm (b) -0.633 0.073 -8.694 0.000
## self_stigma ~
## familism (d) -0.151 0.052 -2.924 0.003
## help_seeking_attitudes ~
## familism (e) 0.133 0.062 2.165 0.030
##
## Variances:
## Estimate Std.Err z-value P(>|z|)
## .hlp_skng_tttds 0.910 0.080 11.336 0.000
## .familism 0.963 0.085 11.336 0.000
## .self_stigma 0.668 0.059 11.336 0.000set.seed(230925) #required for reproducible results because lavaan introduces randomness into the calculations
Combined_fit <- lavaan::sem(Combined, data = d, se = "bootstrap", missing = 'fiml', bootstrap = 1000)cFITsum <- lavaan::summary(Combined_fit, standardized = TRUE, rsq=T, ci=TRUE)
cParamEsts <- lavaan::parameterEstimates(Combined_fit, boot.ci.type = "bca.simple", standardized=TRUE)
cFITsum## lavaan 0.6.16 ended normally after 1 iteration
##
## Estimator ML
## Optimization method NLMINB
## Number of model parameters 11
##
## Number of observations 257
## Number of missing patterns 1
##
## Model Test User Model:
##
## Test statistic 91.302
## Degrees of freedom 1
## P-value (Chi-square) 0.000
##
## Parameter Estimates:
##
## Standard errors Bootstrap
## Number of requested bootstrap draws 1000
## Number of successful bootstrap draws 1000
##
## Regressions:
## Estimate Std.Err z-value P(>|z|) ci.lower
## help_seeking_attitudes ~
## publc_stgm (c) -0.231 0.123 -1.883 0.060 -0.479
## familism ~
## publc_stgm (a) -0.176 0.100 -1.760 0.078 -0.377
## help_seeking_attitudes ~
## self_stigm (b) -0.633 0.101 -6.242 0.000 -0.842
## self_stigma ~
## familism (d) -0.151 0.063 -2.379 0.017 -0.287
## help_seeking_attitudes ~
## familism (e) 0.133 0.063 2.118 0.034 0.016
## ci.upper Std.lv Std.all
##
## 0.004 -0.231 -0.137
##
## 0.010 -0.176 -0.118
##
## -0.431 -0.633 -0.468
##
## -0.034 -0.151 -0.179
##
## 0.262 0.133 0.117
##
## Intercepts:
## Estimate Std.Err z-value P(>|z|) ci.lower ci.upper
## .hlp_skng_tttds 7.024 0.366 19.202 0.000 6.258 7.753
## .familism 4.065 0.226 18.022 0.000 3.629 4.506
## .self_stigma 2.872 0.241 11.924 0.000 2.442 3.387
## Std.lv Std.all
## 7.024 6.256
## 4.065 4.114
## 2.872 3.456
##
## Variances:
## Estimate Std.Err z-value P(>|z|) ci.lower ci.upper
## .hlp_skng_tttds 0.910 0.125 7.265 0.000 0.665 1.147
## .familism 0.963 0.085 11.305 0.000 0.786 1.129
## .self_stigma 0.668 0.061 10.906 0.000 0.549 0.790
## Std.lv Std.all
## 0.910 0.722
## 0.963 0.986
## 0.668 0.968
##
## R-Square:
## Estimate
## hlp_skng_tttds 0.278
## familism 0.014
## self_stigma 0.032cParamEsts## lhs op rhs label est se z
## 1 help_seeking_attitudes ~ public_stigma c -0.231 0.123 -1.883
## 2 familism ~ public_stigma a -0.176 0.100 -1.760
## 3 help_seeking_attitudes ~ self_stigma b -0.633 0.101 -6.242
## 4 self_stigma ~ familism d -0.151 0.063 -2.379
## 5 help_seeking_attitudes ~ familism e 0.133 0.063 2.118
## 6 help_seeking_attitudes ~~ help_seeking_attitudes 0.910 0.125 7.265
## 7 familism ~~ familism 0.963 0.085 11.305
## 8 self_stigma ~~ self_stigma 0.668 0.061 10.906
## 9 public_stigma ~~ public_stigma 0.443 0.000 NA
## 10 help_seeking_attitudes ~1 7.024 0.366 19.202
## 11 familism ~1 4.065 0.226 18.022
## 12 self_stigma ~1 2.872 0.241 11.924
## 13 public_stigma ~1 2.216 0.000 NA
## pvalue ci.lower ci.upper std.lv std.all std.nox
## 1 0.060 -0.505 -0.018 -0.231 -0.137 -0.206
## 2 0.078 -0.377 0.010 -0.176 -0.118 -0.178
## 3 0.000 -0.817 -0.422 -0.633 -0.468 -0.468
## 4 0.017 -0.280 -0.024 -0.151 -0.179 -0.179
## 5 0.034 0.011 0.259 0.133 0.117 0.117
## 6 0.000 0.705 1.190 0.910 0.722 0.722
## 7 0.000 0.822 1.173 0.963 0.986 0.986
## 8 0.000 0.565 0.808 0.668 0.968 0.968
## 9 NA 0.443 0.443 0.443 1.000 0.443
## 10 0.000 6.277 7.765 7.024 6.256 6.256
## 11 0.000 3.625 4.505 4.065 4.114 4.114
## 12 0.000 2.432 3.377 2.872 3.456 3.456
## 13 NA 2.216 2.216 2.216 3.332 2.216Visualization
# Calculate mean and standard deviation of familism
mean_familism <- mean(d$familism)
sd_familism <- sd(d$familism)
# Create a new variable indicating high, middle, and low familism
d$familism_group <- cut(
d$familism,
breaks = c(-Inf, mean_familism - sd_familism, mean_familism + sd_familism, Inf),
labels = c("Low", "Middle", "High"),
include.lowest = TRUE
)# Load necessary libraries
library(stats)
library(ggplot2)
# Create a moderation plot using ggplot2
ggplot(d, aes(x = public_stigma, y = self_stigma, color = familism_group)) +
geom_smooth(method = "lm", se = FALSE, size = 1) +
labs(x = "Public Stigma", y = "Self-Stigma") +
scale_color_manual(values = c("#e76f51", "#264653", "#2a9d8f")) + # Adjust colors as needed
theme_minimal() +
ggtitle("Moderating Influence of Familism on Public Stigma and Self-Stigma")## Warning: Using `size` aesthetic for lines was deprecated in ggplot2 3.4.0.
## ℹ Please use `linewidth` instead.
## This warning is displayed once every 8 hours.
## Call `lifecycle::last_lifecycle_warnings()` to see where this warning was
## generated.## `geom_smooth()` using formula = 'y ~ x'