Lecture Two-Sampling

Question 1

What are the indications for a causal factor, when determining the distribution of disease in a population?

Answer 1

The presence of contagious disease
Exposure to a common risk factor (I.e. Lack of vaccination)
Or Both.

Question 2

What is Sampling?

Answer 2

A method to obtain information from a representative subset of a population or group of individuals to make inferences about characteristics of that population or group.

Question 3

What is a census?

Answer 3

When data is collected on all animal of interest. When data is restricted to a subset, it is then termed a sample or study group.

Question 4

Target population?

Answer 4

The population to which the study results are to be extrapolated.

Question 5

Source population?

Answer 5

The immediate population for which the study conclusions are to be used and from which the subset is to be selected (may be a subset of the target population)

Question 6

Sampling frame?

Answer 6

Lists all sampling units in the source population.

Question 7

Sampling units?

Answer 7

Individual members of a sampling frame

Question 8

Internal validity?

Answer 8

Describes the relationship between study (the people actually studied) and source populations.

Question 9

External validity?

Answer 9

Describes the relationship between the source and target populations.

Question 10

If judgement and sampling are correct, what does this indicate?

Answer 10

That external and internal validity are excellent.

Question 11

Non-probability sampling?

Answer 11

NOT based on random sampling techniques. It’s main disadvantage is that ‘representativeness’ of a population can not be quantified, as bias may be present.

Question 12

Types of non-probability sampling?

Answer 12

Convenience sampling-sample is collected because it is easy to obtain.
Judgement sampling-‘representative’ units of the population are selected, what you judge to be important to sample.
Purposive sampling-selection is based on known exposure or disease status (analytic observational studies). Go out with a purpose.

Question 13

Probability sampling?

Answer 13

Based on random sampling techniques. It is a better representative of the study population, with little to no bias, and so is better than non-probability sampling in almost all cases. The assumption is that any of the possible samples from the source population has the same chance of being selected.

Question 14

Types of Probability Sampling?

Answer 14

Simple Random Sample-
Systematic Random Sample-
Stratified Random Sample-
Cluster Sample-
Multi-stage Sampling-

Question 15

Simple Random Sample?

Answer 15

Assigns each sampling unit a number from 1 to N. A sample of n of these units is then chosen using a formal random process (random tables). It is advantageous in that it is a simple concept with every unit having the same probability of being selected (n/N). However, it is not commonly used, as it is not practical and a sampling frame is not often available.

Question 16

What is the important difference between disease in an individual animal and in a population?

Answer 16

An individual with a disease is all or nothing, they have the disease or they don’t. Compared to a population which has ‘x’ amount of disease within it. There is a greater range of clinical signs among the population, ie where one patient may exhibit two of seven symptoms, all may be represented in the population.

Question 17

Systematic Random Sample?

Answer 17

Sample units are listed sequentially at fixed intervals along the sequence.
Sampling Interval = Size of Study Group / Sample Size
The n sampling units are selected from the sampling frame (N) at a regular interval (k). If n (sample size) is fixed, k is the integer nearest N/n (sample frame divided by sample size).
It is advantageous in that it is easily applied in field conditions. However bias may be introduced if the characteristic measured is related to the sampling interval.
If there is periodicity or ordering in the sampling frame, systematic sampling SHOULD NOT BE USED.
Applications include;
- Quality control in production processes in industry.
- Quality control of laboratory results, routine bulk handling of samples.

Question 18

Stratified Random Sample?

Answer 18

Factor(s) which are likely to influence the outcome are identified and used prior to selecting the sample, in order to divide the sampling frame into strata. Then a simple random sample, or systematic random sample is selected within each stratum.
This method improves precision of the estimate, and assures that all strata are included in the sample. It is also more flexible in that there are different sampling percentages in various strata.
It requires prior knowledge of which strata each sampling unit belongs to.

Question 19

Cluster sample?

Answer 19

Groups or clusters of individuals are selected. The selection of these clusters can be by simple random sampling, systematic sampling or stratified sampling. The unit of concern is the animal, but sample units are aggregates of study units. There is random sampling of clusters, and all individuals within a cluster are tested.
It is often used because it is easy to make selections and is practical for field research. However, the information is biased and estimates are less precise, as the occurrence of disease (especially infectious disease), varies more between clusters than within clusters.

Question 20

What does the size of the sample needed depend upon?

Answer 20

The size of the animal population in question.
The likely prevalence if the disease is present. Historical data will be required, but if this is not available, set the prevalence at 50%.
The reliability required of the conclusions.

Question 21

Define Surveillance.

Answer 21

The systematic collection, analysis, interpretation, and dissemination of health data on an ongoing basis, to gain knowledge of the pattern of disease occurrence and potential in a population, in order to control and prevent disease in the population.
(Looking at food-borne disease across Australia)

Question 22

Difference between surveillance and monitoring?

Answer 22

The difference is that monitoring is an intermittent (regular or irregular) series of observations in time, carried out to show the extent of compliance with a formulated standard or degree of deviation from an expected norm.
(I.e. Herd health programmes)

Question 23

What is the goal of surveillance?

Answer 23

The eradication and control of animal diseases and control of zoonoses.

Question 24

What is a null hypothesis?

Answer 24

There is no association between exposure to a risk factor and the occurrence of disease.

Question 25

What is an Alternative Hypothesis?

Answer 25

There is an association between exposure to a risk factor and the occurrence of disease.

Question 26

What is a Type I error? (alpha)

Answer 26

It is the probability of declaring a difference to be statistically significant when in reality no real difference exists in the population.
Confidence = 1 - alpha

Question 27

What is a Type II error? (beta)

Answer 27

It is the probability of declaring a difference to be statistically non-significant when there is a real difference in the population.
Power = 1 - Beta

Question 28

Important concepts in statistical hypothesis testing?

Answer 28

There is a null hypothesis (H0) which is to be tested with respect to whether it can be rejected or not in favour of an alternative hypothesis (H1).
Then there is a chance variation (= random error) which can result in making either a Type I or Type II error.
A Type I error (= alpha - error, or false positive) refers to incorrectly rejecting the null hypothesis.
The p-value quantifies the probability that the observed association is due to random variation (often set to 0.05). It is the likelihood of a Type I error.
In a Type II error (= Beta - error, or false negative), the null hypothesis is false but not rejected.

Question 29

How are the effect of treatment and effect of control related for;

a) The null hypothesis (H0)
b) One-tailed hypothesis test
c) Two-tailed hypothesis test

Answer 29

a) null hypothesis; H0: effect of treatment = effect of control
b) one-tailed hypothesis test; H1: effect of treatment > effect of control.
OR
effect of treatment

Question 30

Symbols represent?
N?
d? 
n?
Alpha?
k?
Z?
p?
L?

Answer 30

N; sampling frame
n; sampling units
Alpha; 1 - desired confidence level
k; sampling interval
Z; value of standard normal distribution corresponding to the desired confidence level (e.g. 1.96 for 95% confidence)
p; expected proportion of diseased animals in the population
L; decision precision of the prevalence estimate