Methodology
- Content Analysis
- A research method that organizes words, themes, or concepts within some given qualitative data, typically text. Researchers quantify specific words, themes, or concepts and analyze the data to reveal the presence of hidden elements, new relationships, and more rigorous understanding of colloquialisms.
- Difference in Difference Estimation
- The method attempts to capture the effects of an exogenous change or treatment on a population. The model identifies a control group with similar characteristics to the treatment group that has not been exposed to the treatment. The outcome of the control group serves as a proxy for the treated group if they had not been treated. Investigators use the difference between the average post-treatment outcomes between the treatment and control groups to determine the treatment effects.
- K-means Clustering Approach
- The distance-based algorithm creates clusters of data then evaluates various combinations. The final result minimizes the sum of distances between data points and their cluster centroid.
- Clustering Large Applications (CLARA) Algorithm: A clustering technique used to deal with more than several thousand objects. Modelers can decrease computation time and memory storage issues by employing the Partitioning Around Medoids (PAM) algorithm.
- Logistic Multi-Level Regression Models
- The regression estimates the chance that an event will happen while considering data dependency. The model estimates the likelihood as a function of lower-level variables, higher-level variables, and cross-level interactions.
- Mixed Methods
- A research method that gathers and analyzes both qualitative and quantitative data within the same research project.
- Multinomial Logistic Regression
- The nominal outcome variable model estimates the log odds of outcomes as a linear combination of the independent variables.
- Propensity Score Matching (PSM) on Treatment and Control
- The matching technique estimates the assignment probability of treatment, based on observed baseline characteristics. The process allows investigators to mimic some characteristics of a randomized controlled trial
- Treatment Group
- The population subject to a change in environment
- Control Group
- A population similar to the treatment group in characteristic but not exposed to the change in environment.