Troubleshooting Guide and FAQ for jjstatsplot

Troubleshooting Guide and FAQ

This comprehensive guide addresses common issues, error messages, and frequently asked questions when using jjstatsplot in both jamovi and R.

Installation Issues

Q: jjstatsplot doesn’t appear in jamovi’s module list

Solutions: 1. Check jamovi version: Requires jamovi ≥ 1.2.19 Help → About jamovi (check version number)

2. Refresh module library:
   • Close jamovi completely
   • Restart jamovi
   • Try searching again in jamovi library
3. Manual installation:
   • Download .jmo file from GitHub releases
   • jamovi → Modules → Install from file
4. Internet connection: Ensure stable internet for library access

Q: “Module failed to install” error

Solutions: 1. Clear module cache: - Close jamovi - Delete jamovi modules folder (varies by OS) - Restart jamovi and reinstall

2. Permission issues: Run jamovi as administrator (Windows) or check write permissions

3. Dependency conflicts: Uninstall conflicting modules temporarily

Q: R installation fails with dependency errors

# Common dependency installation issues:

# Issue: Package compilation fails
# Solution: Install binary packages
install.packages("jjstatsplot", type = "binary")

# Issue: ggstatsplot not found
# Solution: Install dependencies first
install.packages(c("ggstatsplot", "jmvcore", "R6"))
devtools::install_github("sbalci/jjstatsplot")

# Issue: devtools not available
# Solution: Install devtools first
install.packages("devtools")

Data-Related Issues

Q: “No data to plot” error

Common Causes & Solutions:

1. Missing variable selection:

   • jamovi: Ensure variables are dragged to appropriate boxes
   • R: Check variable names are correct and exist in data

2. All missing values:

   # Check for missing data
   summary(your_data)
   
   # Remove rows with missing values for specific variables
   clean_data <- your_data[complete.cases(your_data[c("var1", "var2")]), ]

3. Incorrect variable types:

   # Check variable types
   str(your_data)
   
   # Convert variable types if needed
   your_data$categorical_var <- factor(your_data$categorical_var)
   your_data$numeric_var <- as.numeric(your_data$numeric_var)

Q: “Insufficient data for analysis” message

Solutions: 1. Check sample sizes: - Minimum 10 observations for most analyses - At least 5 observations per group for group comparisons - Correlation analyses need at least 10 complete pairs

2. Group-specific issues:

   # Check group sizes
   table(your_data$grouping_variable)
   
   # Remove small groups if necessary
   group_counts <- table(your_data$grouping_variable)
   large_groups <- names(group_counts[group_counts >= 5])
   filtered_data <- your_data[your_data$grouping_variable %in% large_groups, ]

Q: Variables show as “character” instead of “factor”

Solution:

# In R: Convert character to factor
your_data$category <- factor(your_data$category)

# In jamovi: 
# 1. Double-click variable name in data view
# 2. Change "Data Type" from "Text" to "Nominal" or "Ordinal"
# 3. Set appropriate factor levels if needed

Statistical Test Issues

Q: “Statistical test failed” or “Assumptions violated” warnings

Common Issues & Solutions:

1. Normality violations:

   # Check normality
   shapiro.test(your_data$variable)
   
   # Solutions:
   # 1. Use non-parametric alternatives (usually automatic in jjstatsplot)
   # 2. Transform data
   log_transformed <- log(your_data$positive_variable)
   sqrt_transformed <- sqrt(your_data$variable)
   
   # 3. Use robust statistical methods (available in options)

2. Homogeneity of variance violations:

   # Check variance homogeneity
   bartlett.test(variable ~ group, data = your_data)
   
   # Solutions:
   # 1. Use Welch's t-test (often default)
   # 2. Use non-parametric tests
   # 3. Transform data to stabilize variance

3. Small sample sizes for specific tests:

   • Fisher’s exact test for small contingency tables
   • Bootstrap methods for small samples
   • Non-parametric alternatives

Q: Correlation analysis produces NaN or infinite values

Solutions:

# Check for issues
summary(your_data[, numeric_variables])

# Remove constant variables
non_constant <- sapply(your_data[numeric_variables], function(x) length(unique(x)) > 1)
valid_variables <- numeric_variables[non_constant]

# Check for extreme outliers
outlier_check <- function(x) {
  Q1 <- quantile(x, 0.25, na.rm = TRUE)
  Q3 <- quantile(x, 0.75, na.rm = TRUE)
  IQR <- Q3 - Q1
  extreme_outliers <- x < (Q1 - 3 * IQR) | x > (Q3 + 3 * IQR)
  return(sum(extreme_outliers, na.rm = TRUE))
}

sapply(your_data[valid_variables], outlier_check)

Plot Display Issues

Q: Plots appear blank or don’t display

Solutions:

1. jamovi-specific:

   • Refresh analysis (click away and back)
   • Resize plot area
   • Check if plot is in “plot2” section for grouped analyses

2. R-specific:

   # Check if plot object exists
   result <- jjhistostats(data = mtcars, dep = "mpg", grvar = NULL)
   
   # For single plots
   if (exists("result") && !is.null(result$plot)) {
     print(result$plot)
   }
   
   # For grouped plots
   if (exists("result") && !is.null(result$plot2)) {
     print(result$plot2)
   }

3. Graphics device issues:

   # Reset graphics device
   dev.off()
   
   # For RStudio users
   # Go to Plots pane → Export → Copy to Clipboard

Q: Plot text is too small or overlapping

Solutions:

1. In jamovi:

   • Increase plot size in results panel
   • Export plot at higher resolution

2. In R:

   # Adjust plot size
   result <- jjhistostats(data = mtcars, dep = "mpg", grvar = NULL)
   
   # Modify plot theme
   improved_plot <- result$plot + 
     ggplot2::theme(
       text = ggplot2::element_text(size = 12),
       axis.text = ggplot2::element_text(size = 10),
       plot.title = ggplot2::element_text(size = 14)
     )
   
   print(improved_plot)

Q: Statistical annotations are cut off or overlapping

Solutions:

# Adjust plot margins
library(ggplot2)

result <- jjscatterstats(data = mtcars, dep = "mpg", group = "hp")

# Increase plot margins
fixed_plot <- result$plot + 
  theme(
    plot.margin = margin(t = 20, r = 20, b = 20, l = 20, unit = "pt")
  )

# For very long annotations, consider:
# 1. Shorter variable names
# 2. Custom subtitle/caption
# 3. Separate statistical table

Performance Issues

Q: Analysis takes very long time or jamovi freezes

Solutions:

1. Large datasets:

   # Sample large datasets
   if (nrow(your_data) > 10000) {
     sampled_data <- your_data[sample(nrow(your_data), 5000), ]
   }
   
   # Or focus on specific subsets
   subset_data <- your_data[your_data$important_condition == "focus_group", ]

2. Complex grouping:

   • Reduce number of groups
   • Combine similar categories
   • Use simpler analyses for exploration

3. Memory issues:

   # Check memory usage
   object.size(your_data)
   
   # Clean workspace
   rm(list = ls()[!ls() %in% c("essential_objects")])
   gc()  # Garbage collection

Common Error Messages

“Error in eval(expr, envir, enclos): object ‘variable_name’ not found”

Cause: Variable name mismatch or typo

Solutions:

# Check available variables
names(your_data)

# Check for exact spelling
"variable_name" %in% names(your_data)

# Common issues:
# - Case sensitivity: "Variable" vs "variable"
# - Spaces: "Variable Name" vs "Variable_Name"
# - Special characters: accents, symbols

“Error: Aesthetics must be either length 1 or the same as the data”

Cause: Data length mismatch, often from grouping issues

Solutions:

# Check data consistency
table(your_data$grouping_variable, useNA = "always")

# Remove missing values in grouping variable
clean_data <- your_data[!is.na(your_data$grouping_variable), ]

# Ensure balanced groups if needed
balanced_data <- your_data %>%
  group_by(grouping_variable) %>%
  filter(n() >= 5) %>%
  ungroup()

“Warning: Computation failed in stat_*(): Non-numeric variable”

Cause: Numeric variable treated as categorical or vice versa

Solutions:

# Check variable type
class(your_data$problematic_variable)

# Convert to appropriate type
your_data$numeric_var <- as.numeric(as.character(your_data$numeric_var))
your_data$factor_var <- factor(your_data$factor_var)

# Handle factor levels that look numeric
your_data$category <- factor(your_data$category, 
                            levels = c("Low", "Medium", "High"))

Best Practices to Avoid Issues

Data Preparation

# 1. Always check data structure first
str(your_data)
summary(your_data)

# 2. Handle missing values explicitly
your_data_clean <- your_data %>%
  filter(!is.na(dependent_variable) & !is.na(grouping_variable))

# 3. Set appropriate variable types
your_data$categorical <- factor(your_data$categorical)
your_data$ordered_cat <- ordered(your_data$ordered_cat, 
                                levels = c("Low", "Med", "High"))

# 4. Check for outliers
boxplot(your_data$continuous_variable)

# 5. Verify sufficient sample sizes
table(your_data$grouping_variable)

Analysis Workflow

1. Start simple: Begin with basic plots before adding grouping
2. Check assumptions: Use diagnostic plots and tests
3. Validate results: Compare with alternative methods
4. Document choices: Keep track of data transformations and exclusions

Reporting Guidelines

1. Report sample sizes: Always include n for each group
2. Describe missing data: How were missing values handled?
3. State assumptions: Were statistical assumptions met?
4. Include effect sizes: Don’t rely only on p-values
5. Show uncertainty: Include confidence intervals when available

Getting Additional Help

jamovi Community Resources

• jamovi Forum: forum.jamovi.org
• jamovi Documentation: docs.jamovi.org
• Video Tutorials: Search “jamovi tutorials” on YouTube

ggstatsplot Resources

• Package Website: indrajeetpatil.github.io/ggstatsplot
• GitHub Issues: For ggstatsplot-specific problems
• Vignettes: Comprehensive function documentation

jjstatsplot Specific

• GitHub Issues: github.com/sbalci/ClinicoPathJamoviModule/issues
• Email Support: drserdarbalci@gmail.com
• Package Documentation: Use ?function_name in R

When Reporting Issues

Include the following information: 1. jamovi version (Help → About) 2. jjstatsplot version 3. Operating system 4. Sample data (if possible) 5. Exact error message 6. Steps to reproduce 7. Screenshots (for jamovi issues)

Creating Minimal Reproducible Examples

# For R issues, create minimal example:
library(ClinicoPath)

# Use built-in data
data(mtcars)

# Minimal code that reproduces the problem
result <- jjhistostats(data = mtcars, dep = "mpg", grvar = NULL)
result$plot  # If this fails, report it

This troubleshooting guide covers the most common issues encountered with jjstatsplot. For complex statistical questions, consider consulting with a statistician or posting detailed questions on statistical forums like Cross Validated (stats.stackexchange.com).