Immediately below this paragraph is a table
with links to the exam questions and answers for all the externals way back to 2013.
Focus on the exams for the last three years. Recognise the type of question e.g This is
a question about..... Know how you are going to set out your working when answering each type of question.
There is a lot of reading, and some writing in the exam. Read the resource carefully
Your first task is to sort out whether this is a question about an experiment, an observational
study, a survey, or a poll
Experiments
In an experiment
subjects are randomly divided into two groups (experimental and control).
Baseline measurements are taken
One group receives a 'treatment' or intervention.
This intervention must be the only difference between the two groups.
All other variables must be controlled
The same measurements are taken after the intervention
If there is a difference between the experimental and the control groups, and if that
difference is unlikely to have arisen by chance, we can conclude that the intervention CAUSED the difference
Experiments are the way we establish cause and effect.
Variables
An experimental design has two variables. Everything else is kept constant. They are
The explanatory or independent variable. This is the one that the experimenter changes e.g fertiliser/no fertiliser
or dose a drug
The response or dependent variable. This is the one that the experimenter measures e.g weight
of the crop or heart rate
Confounding or nuisance variables are things that affect the response which have not been
kept constant. Age and ethnicity are common confounding variables in human experiments
Control Group
In order to determine whether the treatment (independent variable) has an effect it is necessary to have a control
group that does not receive the treatment. For example, if we are investigating whether fertiliser affects the weight of
potatoes grown per hectare then the field may be divided into two, with one half fertilised and one half not. In experiments
involving animals, including humans, we often let the experimental subject (the animal) be its own control. For example,
if we are investigating the effect of a drug on heart rate we will measure the heart rate without the drug in the subjects,
and then again with the drug.
Random Allocation
It is very important that experimental subjects are randomly allocated to treatment or control groups.
Observational Studies
Surveys
Samples
Sampling and Non-Sampling Error
Margin of Error
Q1a from the 2021 exam
The rule of thumb is the the margin of error equals 1/√n
1/Sqrt;883 = 0.0337 = 3.37%
1a(ii) Explain why the margin of error is needed to interpret any claims made from this survey
Achieved - The margin of error is needed to take into account the natural variation we
would expect to see from survey to survey just through the process of sampling.
Merit - applies the 'achieved' answer to this survey - For the Federated Farmers Rural Connectivity Survey
2020, every time a
sample of farmers are surveyed and asked the questions on connectivity, we
would expect slightly different results.
