Fig. 2

Scatter plot of the causal relationships between Years of education and outcomes using different MR methods. a Scatter plot of the causal relationships between Years of education and BMD; b Scatter plot of the causal relationships between Years of education and Whole body fat mass; c Scatter plot of the causal relationships between Years of education and Physical activity; d Scatter plot of the causal relationships between Years of education and Time for vigorous physical activity; e Scatter plot of the causal relationships between Years of education and Calcium consumption; f Scatter plot of the causal relationships between Years of education and Sunbath; g Scatter plot of the causal relationships between Years of education and Salt consumption; h Scatter plot of the causal relationships between Years of education and Smoking; i Scatter plot of the causal relationships between Years of education and Coffee intake; j Scatter plot of the causal relationships between Years of education and Fizzy drink; k Scatter plot of the causal relationships between Years of education and Falling risk. The slope of each line corresponds to the causal estimates for each method. The individual SNP effect on the outcome (point and vertical line) against its effect on the exposure (point and horizontal line) was delineated in the background. BMD, bone mineral density; MR, Mendelian randomization; SNP, single nucleotide polymorphism