Change Trajectories of Couples Therapy

Different couples can exhibit unique change patterns over the course of a couples therapy. This project explores the linear and nonlinear trajectories in couples therapy by leveraging changepoint package by Killick et al. (2014). You can compare the goodness-of-fit for linear and nonlinear modeling of the change trajectories for couples.

This project includes following functions:

1. Find the critical change points in the routine outcome monitoring of each partner
2. Compare the goodness-of-fit for linear, loglinear, quadratic, cubic, and changepoint models using the McFadden's pseudo R-squared
3. Examine the similarity of the change patterns between the partners to see if they change in synchrony. Considering the lead-lag effects, you can also specify the time delay for up to 2 time points (see config.R for more details)

Copy the repository

git clone https://github.com/XinY-Z/chg_traj_couple

Installation

install.packages("data.table")
install.packages("changepoint")

Usage

First, open the config.R file and specify 'coupleid, partnerid, and outcome variables. These variables should match the couple id, partner id, and outcome variable in your dataset. Save it and run the following command in R console.

source("run.R")
# When you run the script, you will be prompted to enter the file path of your data.

Data

The data should be in a long format with the following columns:

• coupleid: Unique couple identifier
• partnerid: Unique partner identifier
• time: Time point of the measurement
• outcome: Outcome variable of interest

Example

# Customize configuration
source("config.R")
config$coupleid <- <your couple id variable>
config$partnerid <- <your partner id variable>
config$outcome <- <your outcome variable>
# save to config.R

# Run the script
source("run.R")

# Enter the file path of the data
> data/example_data.csv

# This will return three dataframes

# Critical Change Points
   coupleid partnerid change_point
1         1         1          3,5
2         1         2          2,3
3         2         1           
4         2         2      6,12,16

# McFadden's Pseudo R-squared
   coupleid partnerid changepoint linear loglinear  quad  cube
1         1         1       0.000  0.031     0.099 0.016 0.013
2         1         2       0.000  0.109     0.203 0.074 0.059
3         2         1       0.196  0.107     0.080 0.097 0.075
4         2         2       0.145  0.018     0.054 0.005 0.002

# Similarity of Change Patterns
   coupleid similarity    type
1         1      0.000 jaccard
2         2      0.000 jaccard
3         3      0.167 jaccard

License

This project is licensed under the MIT License - see the LICENSE.md file for details.

Acknowledgements

• Killick, R., & Eckley, I. A. (2014). changepoint: An R package for changepoint analysis. Journal of Statistical Software, 58(3). https://doi.org/10.18637/jss.v058.i03

Citation

If you use this project in your research, please cite the following paper:

@unpublished{zhang2023,
  title={Comparing Change Trajectories in Couples Therapy: A Data-Driven Approach},
  author={Zhang, Xinyao and Baucom, Brian R. W.},
  year={2023}
}