RMS composite metrics - exploratory data analysis: Module 6

(update 2024-11-10)

Note: All analyses are done by round, using all respondents.

1. Correlation among select composite metrics across modules

Note:
* Varying colors show direction of coefficients (also printed). See legend.
* Blank color means the correlation coefficient is not significant with p-value >=0.1.
* Indices are in alphabetical order for comparison purposes across rounds.

Metrics for all women/households

Round 1

Round 2

Round 3

2. Assessment individual composite metrics by module

For each index metrics, we present the following:
1. Mean values of individual items by round
2. Correlation matrix
3. PCA results
* Percent variation explained by component/dimension
* Bi-plot of individual items with respect to the first and second components

Analysis by round allows us to think if there are seasonal/temporal differences in how the individual items are correlated. Of note, even if each round data are similar, interpretation is limited to the study setting. But, still a good contribution….

Module 6: Individuial or household resilience

Metric: indris

List of composite metrics
variable_name	construction	item_type	number_of_items	possible_minimum_score	possible_maximum_score
indris	simple average of items	5 categories (1-5)	6	1	5

Summary by round

Correlation matrix by round

Round 1

PCA by round

Distribution of the composite measure by round

Metric: hhrisris

List of composite metrics
variable_name	construction	item_type	number_of_items	possible_minimum_score	possible_maximum_score
hhris	simple average of items	5 categories (1-5)	9	1	5