Table 3: Longitudinal analysis: living arrangement and loneliness on psychological well-being at followup.

Source of variationDepressive symptoms scores at followup ( 𝑁 = 1 8 4 1 )SF-12_MCS scores at followup ( 𝑁 = 1 7 9 0 )

dfMean squareF 𝑃 𝑅 2 dfMean squareF 𝑃 𝑅 2

Base model:
 Age125.859.340.021388.2111.020.001
 Gender (0 = male, 1 = female)113.464.860.0319.740.280.60
 Race10.440.160.6911.250.040.85
 Education12.380.860.3511.410.040.84
 Marital status14.401.590.2119.080.260.61
 Number of medical problems135.3912.78<0.0011702.4519.95<0.001
 Baseline GDS1788.88284.87<0.001
 Baseline ADL139.7614.36<0.0011340.1799.660.002
 Baseline MMSE10.890.320.571107.093.040.08
 Social contact frequency score18.403.030.081121.463.450.06
 Social activities score121.107.620.011447.6012.71<0.001
 Production activities score135.3212.75<0.0011156.464.440.04
 Fitness activities score14.701.700.1910.060.020.97
 Health activities score18.743.160.080.2201169.034.800.03
 Baseline SF-12_MCS well-being14179.08118.67<0.0010.140
Model 1a: plus living alone versus with others (1,0) only15.301.910.170.2201130.543.710.050.143
Model 1b: plus Lonely versus not lonely (1,0) only137.6714.93<0.0010.2231578.9417.14<0.0010.145
Model 2: plus living alone and loneliness (main effects)
 Living alone versus with others (1,0)14.631.840.18165.101.930.17
 Lonely versus not lonely (1,0)135.0713.91<0.0010.2211526.4815.58<0.0010.146
Model 3: Main effects and interaction
 Living alone versus with others (1,0)18.563.400.071152.814.530.03
 Lonely versus not lonely (1,0)132.5712.92<0.0011560.4216.61<0.001
 Living arrangements*Loneliness14.391.740.190.2421110.653.280.070.158