Table 4 Models predicting personal measurements using outdoor ambient exposure estimates and controlling for repeated measures on subjects
Model description (random and fixed effects)Variance component (95% confidence limits)% Variance explained* (compared to baseline)
Within subject (σWS) (temporal)Between subject (σBS) (spatial)σWSσBSTotal†
NO personal (dependent)
    Baseline (subject only)0.332 (0.242 to 0.485)0.188 (0.101 to 0.470)
        +LUR NO home0.214 (0.156 to 0.312)0.180 (0.107 to 0.366)36424
        +LUR NO home+work0.210 (0.153 to 0.306)0.151 (0.086 to 0.327)372031
        +Ambient monitors IDW NO0.208 (0.152 to 0.304)0.162 (0.094 to 0.343)371429
NO2 personal (dependent)
    Baseline (subject only)0.087 (0.063 to 0.126)0.112 (0.072 to 0.202)
        +LUR NO2 home (annual)0.086 (0.062 to 0.125)0.110 (0.070 to 0.199)122
        +LUR NO2 home+work (annual)0.084 (0.061 to 0.122)0.104 (0.066 to 0.190)376
        +Ambient monitors IDW NO2
Absorbance personal (dependent)
    Baseline (subject only)0.165 (0.118 to 0.246)0.025 (0.006 to 1.341)
        +LUR absorbance home
        +LUR absorbance home+work
        +Ambient monitors IDW PM2.50.146 (0.105 to 0.219)0.029 (0.009 to 0.422)11(–)8
PM2.5 personal (dependent)
    Baseline (subject only)0.169 (0.121 to 0.251)0.060 (0.026 to 0.251)
        +LUR PM2.5 home
        +LUR PM2.5 home+work
        +Ambient monitors IDW PM2.50.154 (0.110 to 0.230)0.075 (0.036 to 0.230)9(–)
  • *The “% variance explained” is simply a per cent change between the variance in the baseline model compared to one with an exposure estimate; “(–)” means that less variance was explained than by the baseline model.

  • †Per cent of variance explained from significant models (between- and within-subject) represents the amount of spatial and temporal variability, respectively, in personal measurements explained by the model effects.