Figure 1.

Structural equation model on associations between anticipated consequences of an HIV infection, perceived HIV severity and sexual risk behaviour among 1,072 HIV‐negative and never‐tested MSM living in the Netherlands.

The numbers next to each pathway indicate the regression coefficient (β). The β regression coefficients generally represent the change in modelled outcomes (i.e. ending vertex of the path) per 1 unit increase in the independent variable (i.e. the starting vertex of the path). For example the β between the latent variables and expected severity represents the change in severity level per 1 standard deviation increased in the latent variable.