1 - Experimental Design Flashcards
Define experiment
Investigation of the relationship between two variables by deliberately producing a change in one of them and observing the change in the other
What is the independent variable
variable which is directly manipulated
independent variables are the ones you change
What is the dependent variable
variable which may consequently change
dependent ones you measure
Describe Observational experiments and what are its strengths and limitations?
Experimental conditions are not manipulated, one or more responses are measured and
compared with one or more ‘environmental’ measurements.
Strengths
- Allows natural relationships to be studied
- Results may provide a useful source of hypotheses for future manipulative experiments.
Limitations
-Cannot prove the factor which is singled out, is responsible for difference in
variable of interest.
-Not usually possible to demonstrate cause and effect.
-Very difficult to control all the variables
Describe Manipulative experiments and what are its strengths and limitations?
Researcher manipulates and controls the experimental conditions to investigate a cause and effect relationship.
Strengths
- This approach allows a causal relationships to be identified
- Laboratory experiments = highly controlled
Limitations
- Lack of generalisability
- Field experiments = can be very expensive, but are more ‘realistic’
What is the Null Hypothesis (H0)?
The H0 is the commonly accepted fact
Researchers work to reject, nullify or disprove the H0
What is the Alternative Hypothesis (HA)?
The alternative hypothesis (HA) is a statement of what a statistical hypothesis test is set up to establish.
What is the Experimental triad?
Chance, Bias, Confounding
What is confounding?
Where an association between an exposure and an outcome is observed due to the influence of a third variable (a “confounder”)
What is bias?
Refers to how far an ‘average’ statistic lies from the parameter it is estimating, that is, the error which arises when estimating a quantity.
What is the purpose of using a sample of the population?
More practical approach
used to draw inferences about whole population
greater population = closer to true answer
Why are Repetitions useful?
um
Describe Random sampling and its limitations
every sample unit from the population stands an equal chance
Random sampling is preferred as alternatives may lead to bias
Sometimes random sampling is not easy to achieve
Often not possible to sample evenly from the whole population
What is Stratified sampling?
Sampling from a population which can be partitioned
into subpopulations, or strata.
What is Randomised Block Design?
um
