1. FFTFT 2. FTFTF 3. FFFTT 4. TTFFF 5. TFFFF 6. FTFTF 7. FTTFT 8. TFFFF 9. FTFFF 10. TTTFT

Score +1 for a correct answer, -1 for an incorrect answer, and 0 if you did not answer. You can rate your performance as follows:- MBBS (out of 50) below 20: work much harder, 20-25: work harder, 26-32: adequate, 33-40: very good, 41-46: excellent, 47 or more: take over the class, Biomedical Sciences (out of 30) below 12: work much harder, 13-15: work harder, 16-18: adequate, 19-24: very good, 25-27: excellent, 28 or more: take over the class.

1. (a) as large a sample as you like, equivalent to the Normal Distribution test (b) for a different type of data (categorical) (c) as small as two (d) for related samples or a single sample measured twice (e) the differences must be Normal.

2. (a) the numbers of observations can be different, there is no restriction (b,d) we assume that the observations come for two populations which follow Normal distributions with the same mean variance (c) this is what we are testing (e) the test is suitable for small samples, but works with samples of any size.

3. (a) usually has non-zero intercept (b) only if there is no linear relationship whatsoever (c) the slope and intercept have dimensions (d) always goes through the point defined by the means of the two variables (e) if we swap variables we get a different line.

4. (a) these are the minimum and maximum possible values, attained when all points lie on a straight line (b) at least one variable must be from a Normal Distribution (c) should be zero (d) this is for regression, not correlation (e) this is the regression coefficient.

5. (a) 80% of 4 is greater than 3 so all expected frequencies must be greater than 5 (b) for categorical data (c) as b (d) the condition is about the expected frequencies (e) can be as small as 20, if all row and column totals are 10.

6. Blood pressure and height are both continuous variables and we are looking at the relationship between them, so correlation and regression could be used. The paired t test is used to look at measurements of the same thing under two conditions, for example blood pressure before and after eating a meal. The two sample t test is used to look at the difference in mean between two groups, for example the difference in mean blood pressure between medical and biomedical science students. The chi-squared test is for categorical (classified) data, for example the relationship between wearing glasses, contact lenses or neither and eye colour.

7. (a) we can only do this if we can list the population in some way. (b) by standard errors, confidence intervals, etc. (c) The estimates of population values will not be biased. (d) it can be very difficult, because of the need for a list or sampling frame. (e) yes, the list or sampling frame from which it was drawn.

8. (a) cases are subjects with the disease (b, c, d) this is an
observational study *not* a trial (e) this is not a cohort study.

9. (a) we cannot conclude from an observational study that an association
demonstrates causation, so the relationship here is not necessarily causal, it
could be due to other factors or it could be the other way round, lung cancer
causing low cholesterol (cohort studies showed this to be the explanation) (b)
the difference is significant (c) the difference is significant (d) we only
know that they are related, not how closely they are related (e) the
cholesterol was measured *after* the lung cancer has developed, see
(a).

10. (a) the probability is less than 0.0001, which is very small and hence unlikely. (b) in a case-control study, the odds ratio provides an estimate of the relative risk. (c) the confidence interval means that we estimate the population value to be between these limits (d) association does not imply causation, this is an observational study. Smoking and stroke may both be related to some other factor. (e) an increased risk means that they are more likely to have strokes, but does not imply causality.

**Back to MCQ and EMI Self Test Term 2**

This page maintained by Martin Bland.

Last updated: 29 June, 2004