Statistical Test Result #182 — Analysis Report
Contents
- glm (#182) completed
1. Glm
2026-09-21 02:15:12 · statistical_test
The glm on GH_COUNT completed.
| family | poisson |
|---|---|
| aic | 2948.86159 |
| deviance | 1942.210341 |
| null_deviance | 5619.785632 |
| pseudo_r2 | 0.654398 |
| n_obs | 207 |
| dispersion_phi | 9.474405 |
| deviance_gof_p | 0.0 |
| method | glm |
| value_col | GH_COUNT |
| significant | True |
coefficients
| term | coef | se | z | p | ci_lower | ci_upper |
|---|---|---|---|---|---|---|
| const | 4.536949 | 0.019579 | 231.719669 | 0.0 | 4.498574 | 4.575324 |
| ENV_SOILTN | -5.442332 | 0.119154 | -45.674756 | 0.0 | -5.675869 | -5.208794 |
coef forest
fitted vs observed
residuals vs fitted
Warnings
- The data appear overdispersed relative to the Poisson model (Pearson phi = 9.47; values near 1 are expected). Coefficient standard errors are likely too small — refit with the Negative Binomial family.
- The deviance goodness-of-fit test rejects adequate fit (p = 3.402e-280); residual structure may remain unexplained.
Method & Provenance
- Method
- Count GLM (Poisson / Negative Binomial)
- Engine
- Python · statsmodels 0.14.6
- Input dataset
- gallery — catg_example_05_grasshopper_pest_pressure.csv
- Started
- 2026-09-21 02:15:12.174965
- Permanent link
- https://cat-g-omxpert.wishartlab.com/results/by-token/e5dac942c0614cd5bc9baa1859614534
| Parameter | Value |
|---|---|
| test_type | glm |
| value_col | GH_COUNT |
| predictor_cols | ENV_SOILTN |
Command history
# python import statsmodels.api as sm # python model = sm.GLM(y_counts, X_predictors, family=sm.families.Poisson()) # python fit = model.fit()