Experimental Design Flashcards Preview

Quantitative Methods in Health > Experimental Design > Flashcards

Flashcards in Experimental Design Deck (58)
Loading flashcards...
1

What are the three principles of experimental design?

Control Randomise Use enough subjects

2

What is an experimental group?

A group of individuals receiving a treatment whose effect we seek to understand.

3

What two things does an ideal design use?

Both the comparison of two or more treatments, and chance assignment of subjects to treatments

4

What is a control?

A treatment meant to serve as a baseline with which the experimental group is compared.

5

What do observational studies do?

Observe individuals and measure variables of interest but do not attempt to influence the responses.

6

Can we investigate causality in observational studies? Why/Why Not?

In observational studies we are not controlling the conditions, so we cannot investigate any causality issues.

7

What is the best method of determining causation?

To conduct a well-designed experiment.

8

When experiments are not possible, what are three examples of when good evidence of causation exists?

The relationship between the variables is observed in many studies of different types. The association holds when the effects of plausible other variables are taken into account. A plausible scientific explanation exists for a relationship between the variables. (i.e. a lot of circumstantial evidence is needed)

9

When are two variables confounded?

When their effects on a response variable cannot be distinguished from each other.

10

What characterises a single-blind study?

In a single‐blind study, the subject does not know which treatment they are receiving.

11

What characterises a double-blind study?

In a double‐blind study, neither the subject nor the experimenter knows which treatment is being used until all measurements are taken.

12

What does the blinding of subjects eliminate?

The placebo effect.

13

What does blinding the experimenter prevent?

Subjects in one group from being treated differently to those in another.

14

What is a block?

A group of individuals that are known before the experiment to be similar in some way that is expected to affect the response to the treatments.

15

What occurs in a block design?

The random assignment of individuals to treatments is carried out separately within each block.

16

What is an example of a block design?

Matched pairs

17

What does the assessment of the quality of a study design require?

Knowledge of the factors that influence or cause variation in the measured outcomes.

18

What is internal validity?

When conclusions can be appropriately drawn from within the study about the relationship between the explanatory and response variable.

19

What are two questions to ask to assess internal validity?

How well was an experiment done to avoid confounding? Did it measure what it set out to measure?

20

What is external validity?

When conclusions from the study can be appropriately generalised to a wider situation of interest.

21

How does replication allow for interval validity?

It allows the variability due to the uncontrolled factors to be measured so that relationship between explanatory and response variables can be properly assessed.

22

How do controlling variables and blocking allow for interval validity?

By minimizing the effects of random variation.

23

How does randomisation allow for interval validity?

To make sure similar groups are compared. Check to confirm similarity.

24

What are two ways to ensure external validity?

Sampling appropriately with respect to the study material and conditions Replicating a broader range of conditions

25

What is the effect size for a research question?

A measure of how much the truth differs from chance or from a control condition.

26

What is the formula for the effect size of a single population mean?

d=(µ-µ0)/σ

27

What is the formula for the effect size of a single sample mean?

dhat=(x̄-µ0)s

28

What is the formula for the effect size when comparing two population means?

d=(µ12)/σ​​

29

What is the formula for the effect size when comparing two sample means?

dhat=(x̄1-x̄2)/s

30

What is the |d| value for small effect size?

0.2