Rejection of the SHPM hypothesis by fitting a generalized log-log linear regression model (GLMloglog) to a limiting dilution transplantation assay (LDTA) performed with whole splenic lymphomatous cells (SLCs)
| Characteristic . | Definition . |
|---|---|
| Cell population | SLCs |
| Deviance dispersion statistic* | 1.89 |
| β | 0.473 |
| SE(β) | 0.159 |
| Null hypothesis, β = 1 | |
| P(z)† | 0.0009 |
| P(χ2)‡ | 0.0116 |
| SHPM hypothesis | Rejected |
| Characteristic . | Definition . |
|---|---|
| Cell population | SLCs |
| Deviance dispersion statistic* | 1.89 |
| β | 0.473 |
| SE(β) | 0.159 |
| Null hypothesis, β = 1 | |
| P(z)† | 0.0009 |
| P(χ2)‡ | 0.0116 |
| SHPM hypothesis | Rejected |
The deviance dispersion statistic can be considered as a summary measure of goodness of fit of the GLMloglog to the data and should be inferior to 1.5-2 when the model adequately fits the data.5
P(z) associated to the Wald statistic Z.4
P(χ2) associated to the standard likelihood ratio test used to compare nested models.6 The GLMloglog holds, as demonstrated by the rejection of the null hypothesis β = 0 by both Wald and likelihood ratio test statistics with P < .005 (data not shown).