Could differential under-reporting of loneliness between men and women bias the gender-specific association between loneliness duration and rate of memory decline? A probabilistic bias analysis of effect modification. Academic Article uri icon