1.8 Biostatistics 2

Question 1

What are three factors that can effect a studies results?

Answer 1

Bias
Confounding/Effect modification
Chance

Question 2

What are the 7 stages to hypothesis testing?

Answer 2

Research question
Sample and conduct study
Null & Alternative hypothesis
Identify level of significance / probability
Calculate the test statistic
Obtain p-value / CI
Interpret and make conclusions

Question 3

What three things does a research question need to address?

Answer 3

Identify population of interest
Define outcome of interest
Define factors of interest

Question 4

What is the relationship between sample size and the mean

Answer 4

The larger the sample size the greater the chances that the mean and SD will be the true mean and SD of the population.

Question 5

What does increasing the sample size of a study do?

Answer 5

Reduces random error, increases precision and increases power

Question 6

Why is the sample size calculated prior to the study?

Answer 6

To ensure you have the required sample size to observe an effect/association.

Question 7

What does the Null Hypothesis state?

Answer 7

Postulates that, in the population, there is no difference between groups and that any difference found is therefore due to chance/sampling variation.

Question 8

What does the Research (Alternative) Hypothesis state?

Answer 8

That in the population there is a difference between groups. The difference will be two tailed if it doesn’t specify which way the difference will fall.

Question 9

Why will there always be random error in the sample?

Answer 9

Because we are not testing the whole population

Question 10

What is the alpha value?

Answer 10

The probability that we will obtain an extreme test statistic assuming that the null hypothesis is true. Represents threshold beyond which the null hypothesis will be rejected.

Question 11

What value is the alpha value usually set at?

Answer 11

Convention is to set it at 0.05. i.e. We are prepared to make an error 5% of the time.

Question 12

What is a Type 1 error?

Answer 12

Rejecting the Null Hypothesis (there is a difference between groups) when in fact the Null Hypothesis is true (no difference). Equal to the alpha value.

Question 13

What is a Type 2 Error?

Answer 13

Accepting the Null Hypothesis (no difference) when in fact the Null hypothesis is not true (there is no difference between groups). Equal to the beta value.

Question 14

How are Type 1 errors reduced?

Answer 14

By reducing the alpha value, i.e. from 0.05 to 0.01. Though this increases the chance of a Type II error.

Question 15

What is Power?

Answer 15

It is the probability of finding an association (result) in our sample if there is a true association in the population.

Question 16

What do we do if the p-value is <0.05

Answer 16

Reject the null hypothesis. Accept there is a difference between the two means and that they do not come from the same population.

Question 17

What is the 95% Confidence Interval?

Answer 17

It is the probability that the sample mean lies within 1.96 standard errors of the population mean.
If we drew multiple random samples from our population of interest, 95% of the confidence intervals would contain the true population mean.

Question 18

What does a wide Confidence Interval indicate?

Answer 18

Poor precision and more random error

Question 19

What does a narrow Confidence Interval indicate?

Answer 19

Good precision and little random error

Question 20

What is the null value for mean differences and proportions?

Answer 20

0

Question 21

What is the null value for risk rations and odds ratios?

Answer 21

1

Question 22

What do we do if the 95% CI includes the null value?

Answer 22

p>0.05, cannot reject the null hypothesis

Question 23

What do we do if the 95% CI does not include the null value?

Answer 23

p<0.05, can reject the null hypothesis