With such a sizeable sample size a type I error is possible, in such cases the magnitude of differences becomes more important than the statistical significance.

Both the correlation analysis and the multiple linear regression analysis have very low levels of significance and prediction. Note that the highest recorded r is 0.203 and is in relation to C-peptide. The r with HbA1c is -0.166 which does not provide much confidence in terms of prediction or correlation, specially since we may be under the effect of sample size (type I error as mentioned above). The same is also true for the multiple linear regression, in which the β values obtained are too low to be able to draw any conclusions form this statistical analysis. 横断面研究，样本量是2172，相关性分析和回归分析都有显著性，r值-0.13，beta值-0.019,审稿人认为r和beta过小，怎么回复比较好？