Although the use of short self-report measures is common practice in biobank initiatives, such a phenotyping strategy is inherently prone to reporting errors. To explore challenges related to self-report errors, we first derived a reporting error score in the UK Biobank (UKBB; n = 73,127), capturing inconsistent self-reporting in time-invariant phenotypes across multiple measurement occasions. We then performed genome-wide scans on the reporting error score, applied downstream analyses (linkage disequilibrium score regression and Mendelian randomization) and compared its properties to the UKBB participation propensity. Finally, we improved phenotype resolution for 24 measures and inspected the changes in genomic findings. We found that reporting error was present across all 33 assessed self-report measures, with repeatability levels as low as 47% (childhood body size). Reporting error was not independent from UKBB participation, evidenced by the negative genetic correlation between the two outcomes (r_g = −0.77), their shared causes (for example, education) and the loss in self-report accuracy following participation bias correction. Across all analyses, the impact of reporting error ranged from reduced power (for example, for gene discovery) to biased estimates (for example, if present in the exposure variable) and attenuation of genome-wide quantities (for example, 21% relative attenuation in SNP heritability for childhood height). Our findings highlight that both self-report accuracy and selective participation are competing biases and sources of poor reproducibility for biobank-scale research.

尽管使用简短的自我报告措施是生物样本库计划中的常见做法，但这种表型分析策略本身就容易报告错误。为了探索与自我报告错误相关的挑战，我们首先在英国生物样本库 （UKBB;n = 73,127），在多个测量场合捕获时间不变表型的不一致自我报告。然后，我们对报告错误评分进行了全基因组扫描，应用下游分析 （连锁不平衡评分回归和孟德尔随机化），并将其特性与 UKBB 参与倾向进行了比较。最后，我们提高了 24 项措施的表型分辨率，并检查了基因组结果的变化。我们发现所有 33 项评估的自我报告措施都存在报告错误，重复性水平低至 47% （儿童体型）。报告误差并非独立于 UKBB 参与，两个结果之间的负遗传相关性 （r_g = -0.77）、它们的共同原因（例如，教育）以及参与偏倚校正后自我报告准确性的损失证明了这一点。在所有分析中，报告错误的影响范围从降低的功效（例如，用于基因发现）到有偏差的估计（例如，如果存在暴露变量）和全基因组数量的衰减（例如，SNP 遗传力对儿童身高的相对衰减 21%）。我们的研究结果强调，自我报告的准确性和选择性参与都是生物样本库规模研究的竞争性偏倚和可重复性差的根源。