Chapter 21 Multiple choice questions

* not completed

Which tests are appropriate in a case control study to determine whether a genetic variant increases the risk of schizophrenia? Select all correct answers.

Logistic regression correct incorrect

Linear regression correct incorrect

x² test correct incorrect

Fisher's exact test correct incorrect

* not completed

At a locus with two alleles, how many of each allele are present in a sample with the following genotypes: 20 individuals homozygous for one allele, 25 homozygous for the other allele and 40 individuals who are heterozygotes Select correct answer.

Impossible to calculate when the sample is not in Hardy Weinberg equilibrium correct incorrect

60 of one allele, and 65 of the other correct incorrect

80 of one allele and 90 of the other correct incorrect

120 of one allele, and 130 of the other correct incorrect

* not completed

What are the advantages of using a Cochran–Armitage trend test? Select all correct answers.

It is robust to departures from Hardy-Weinberg equilibrium correct incorrect

It allows the specification of different genetic models correct incorrect

It increases power to detect a genetic effect correct incorrect

It requires fewer degrees of freedom than other tests correct incorrect

* not completed

What tests are appropriate for detecting a genetic effect on a quantitative personality measure at a locus with two alleles? Select correct answer.

Test of the difference between two means (a t-test) correct incorrect

x ² test correct incorrect

Linear regression correct incorrect

Logistic regression correct incorrect

* not completed

What P-value threshold is appropriate when testing for a genetic effect of a single locus with two alleles on ten different phenotypes? Select correct answer.

P < 0.05 correct incorrect

P < 0.01 correct incorrect

P < 0.005 correct incorrect

P < 0.0005 correct incorrect

* not completed

Which of the following are examples of P-hacking? Select all correct answers.

Using a threshold of P < 0.05 to determine significance when testing the association of 100 genetic markers with one phenotype correct incorrect

Using four different statistical tests to determine the presence of genetic association and reporting only the result with the lowest P-value correct incorrect

Continuing to repeat an experiment until one result is found where P < 0.05 correct incorrect

In a genetic association test, examining the sexes together, then test them separately and declare a finding of P = 0.02 significant because it is less than the 5% threshold of 0.05 correct incorrect

* not completed

What factors determine the power of a genetic association study for schizophrenia? Select all correct answers.

The sample size correct incorrect

The genetic effect at a locus correct incorrect

The allele frequency of a locus correct incorrect

Disease prevalence correct incorrect

* not completed

What are the consequences of carrying out a genetic association test in an underpowered sample? Select all correct answers.

It means that a finding with P-value <0.05 is untrue correct incorrect

It reduces the chances of discovering a true effect correct incorrect

It overestimates the genetic effect correct incorrect

It means that the P-values obtained from any statistical test must be corrected to take into account the lower power correct incorrect

* not completed

According to Fisher, what does a P-value less than 0.05 mean? Select correct answer.

It means that the hypothesis tested is true correct incorrect

It strongly indicates that the hypothesis tested fails to account for the data observed correct incorrect

It is likely that the hypothesis tested is false correct incorrect

It is the conventional point at which a researcher concludes that the hypothesis tested fails to account for the data observed correct incorrect

* not completed

In Neyman–Pearson hypothesis testing, what does a P-value of less than 0.05 mean? Select correct answer.

The probability of the observed data, or of more extreme data points, is less than 0.05, given that the null hypothesis is true correct incorrect

You can deduce the probability of the experimental hypothesis being true correct incorrect

You know the probability that you are making a wrong decision if you decide to reject the null hypothesis correct incorrect

The null hypothesis can be rejected because it is false correct incorrect

* not completed

Suppose you run a genetic association test between a genetic marker and a personality trait, find a P-value of 0.81, and then notice that the genetic effect goes in different directions in the two sexes, increasing the trait in males (P = 0.002) and decreasing the trait in females (P = 0.01). What do you do? Select all correct answers.

Submit your result to a journal, claiming you have identified a significant gene by sex effect correct incorrect

Repeat your analyses in a separate sample correct incorrect

Ignore the sex result and try to publish the non-significant finding correct incorrect

Contact your colleagues to see if they can help you replicate the gene-by-sex observation correct incorrect

* not completed

What's the difference between type I and type II errors? Select correct answer.

Type I is false acceptance correct incorrect

Type II is false rejection correct incorrect

Type II is worse than Type I correct incorrect

Type I is a false positive while Type II is a false negative correct incorrect

* not completed

What's wrong with setting P <0.05 as a statistical threshold to determine whether a finding should be published or not? Select all correct answers.

P-hacking, publication bias, HARKing and similar practices have devalued the meaning of P < 0.05 correct incorrect

Relying on P-values ignores the importance of reporting effect sizes correct incorrect

A P-value alone doesn't prove anything correct incorrect

P-values are meaningless correct incorrect

* not completed

What is Bayes' theorem? Select all correct answers.

The probability that the evidence supports our hypothesis is the probability of a true finding from all of the positive findings (both true and false) correct incorrect

A research finding is more likely to be true than false if the power multiplied by the prior probability is greater than the statistical threshold correct incorrect

The posterior probability that a hypothesis is true, given new data, is proportional to the product of the likelihood of the new data given the hypothesis and the prior probability of the hypothesis correct incorrect

The probability of the hypothesis being true, given the data correct incorrect

* not completed

Which of the following is true for a genetic association test using Bayesian methods? Select all correct answers.

The significance threshold is an estimate of specificity correct incorrect

Sensitivity is the probability that when there is an association we can detect it (equivalent to the power of a test) correct incorrect

The prevalence of a disease is used to estimate the prior probability of an association correct incorrect

The posterior probability is the P-value of the test correct incorrect

* not completed

Why use Bayesian statistics? Select all correct answers.

Bayesian approaches are simpler to apply than frequentist correct incorrect

Since the data are fixed but the hypothesis may vary, Bayesian approaches don't suffer from the multiple hypothesis testing problem correct incorrect

Bayesian approaches don't struggle to decide what is the most appropriate P-value for a study correct incorrect

Estimating the prior probability of an outcome is always easy to do and justifies using Bayesian approaches correct incorrect