想请老师帮忙解释区分一下survival bias和 selection bias

NO.PZ2016071602000008 问题如下 A tabase of hee funreturns is constructefollows. The first yeof the tabase is 1994. All fun existing of the enof 1994 tha willing to report their verifiereturns for thyeare incluin thyear. The tabase is extenasking the fun for verifiereturns before 1994. Subsequently, fun are aethey are willing to report verifiereturns to the tabase. If a funstops reporting returns, its returns are letefrom the tabase, but the tabase hagreement with fun ththey will keep reporting verifiereturns even if they stop being open to new investors. Whiof the following four statements are correct?I. The tabase suffers from backfilling bias.II. The tabase suffers from survivorship bias.III. The tabase suffers from errors-in-variables bias.IV. The equally weighteannureturn average of funreturns will unr-estimate the performanone woulexpefrom a hee fun A.All the above statements are correct. B.Statements I anII are correct. C.Statements I, II, anIII are correct. Statements II anIV are correct. B is correct. The tabase inclus histories before 1994 antherefore suffers from backfill bias. Next, fun thstop reporting are letefrom the tabase, so it hsurvivorship bias. Errors-in-variables biarises in other contexts, suregression. Finally, the average of funreturns will too high (not too low) because of these two biases. Hence, I. anII. are correct. 如题

NO.PZ2016071602000008 问题如下 A tabase of hee funreturns is constructefollows. The first yeof the tabase is 1994. All fun existing of the enof 1994 tha willing to report their verifiereturns for thyeare incluin thyear. The tabase is extenasking the fun for verifiereturns before 1994. Subsequently, fun are aethey are willing to report verifiereturns to the tabase. If a funstops reporting returns, its returns are letefrom the tabase, but the tabase hagreement with fun ththey will keep reporting verifiereturns even if they stop being open to new investors. Whiof the following four statements are correct?I. The tabase suffers from backfilling bias.II. The tabase suffers from survivorship bias.III. The tabase suffers from errors-in-variables bias.IV. The equally weighteannureturn average of funreturns will unr-estimate the performanone woulexpefrom a hee fun A.All the above statements are correct. B.Statements I anII are correct. C.Statements I, II, anIII are correct. Statements II anIV are correct. B is correct. The tabase inclus histories before 1994 antherefore suffers from backfill bias. Next, fun thstop reporting are letefrom the tabase, so it hsurvivorship bias. Errors-in-variables biarises in other contexts, suregression. Finally, the average of funreturns will too high (not too low) because of these two biases. Hence, I. anII. are correct. 老师这是什么偏差？

NO.PZ2016071602000008 问题如下 A tabase of hee funreturns is constructefollows. The first yeof the tabase is 1994. All fun existing of the enof 1994 tha willing to report their verifiereturns for thyeare incluin thyear. The tabase is extenasking the fun for verifiereturns before 1994. Subsequently, fun are aethey are willing to report verifiereturns to the tabase. If a funstops reporting returns, its returns are letefrom the tabase, but the tabase hagreement with fun ththey will keep reporting verifiereturns even if they stop being open to new investors. Whiof the following four statements are correct?I. The tabase suffers from backfilling bias.II. The tabase suffers from survivorship bias.III. The tabase suffers from errors-in-variables bias.IV. The equally weighteannureturn average of funreturns will unr-estimate the performanone woulexpefrom a hee fun A.All the above statements are correct. B.Statements I anII are correct. C.Statements I, II, anIII are correct. Statements II anIV are correct. B is correct. The tabase inclus histories before 1994 antherefore suffers from backfill bias. Next, fun thstop reporting are letefrom the tabase, so it hsurvivorship bias. Errors-in-variables biarises in other contexts, suregression. Finally, the average of funreturns will too high (not too low) because of these two biases. Hence, I. anII. are correct. 课上说的是插入了个新数据

NO.PZ2016071602000008 关于errors-in-variables bias 1.我看了之后，觉得因为题目中回填等操作使得获取的report return变量其实本身就有偏差（会被高估）了，为什么不属于errors-in-variables bias? 2.所以这个bias主要指在操作上出现的问题，例如漏了变量、模型运算等操作风险事件上吗,。而不是说变量input选的数据集不好？