1
Q
Problem solving cycle - Stage 1 Problem specification and analysis
A
- make the goal of the investigation clear
- plan how to achieve goal
- what data is needed, how it is going to be collected, what will be done with the data
2
Q
Problem solving cycle - Stage 2 Information collection
A
- taking a sample from all possible data (parent population)
- can collect data from whole population (100% sample is called a census)
3
Q
Problem solving cycle - Stage 3 Processing and representation
A
- cleaning the data (dealing with outliers, missing data and errors)
- presenting the data in a diagram that shows its main features
- calculate summary measures e.g. central tendency (mean, standard deviation), spread (range)
4
Q
Problem solving cycle - Stage 4 Interpretation
A
- draw conclusions about investigation
- explain what the calculations tell you, whether the hypothesis fits in
- report observations in words
- ask yourself if findings are realistic, if it answers the question; if not, you have to collect more or different data
5
Q
Sample
A
- provides a set of data values of a random variable
6
Q
Parent population
A
- overarching set can be finite e.g. all professional golf players or infinite e.g. points where a dart can land on a dart board
- Greek letters are used as parameters (measured quantity of a statistical population that summarises or describes an aspect of the population) e.g. μ is the mean
- a sample is representative of the population
7
Q
Sampling frame
A
- a list of the items available to be sampled e.g. sheep in a flock, but some things e.g. cod in the North Atlantic has no frame
8
Q
Sampling fraction
A
- the proportion of the available items that are actually sampled
9
Q
Sampling error
A
- difference between an estimate of a parameter derived from the sample set and the true value
- to reduce sampling error, the sample needs to be as representative as possible of the population
10
Q
Bias
A
- systematic error e.g. using a sample of hockey players to estimate time young women take to run 100m (hockey players are fitter than general population)
11
Q
Sampling techniques - Simple random sampling
A
- every member of the parent population is likely to be selected
- needs sampling frame
- done using random number generator
12
Q
Sampling techniques - Stratified sampling
A
- identify strata (sub-groups) which give different patterns
- ensure all strata are sampled
- proportional stratified sampling - proportional to the size of their populations in the parent population
- selection from each stratum is chosen using simple random sampling
13
Q
Sampling techniques - Cluster sampling
A
- items are chosen from one or several of the strata (now called clusters)
- each cluster should be reasonable representative of the parent population e.g. small cluster of puffin on an island are representative of the entire North European population
14
Q
Sampling techniques - Systematic sampling
A
- choosing individuals from a sampling frame at certain intervals
- cyclic patterns could lead to the drawing of incorrect conclusions
15
Q
Sampling techniques - Quota sampling
A
- choosing individuals for a sample depending on a quota e.g. how many males and females
- non-random version of stratified sampling
16
Q
Sampling techniques - Opportunity sampling
A
- circumstances make a sample readily available
- can result in bias - interviewing delegates of a hospital conference about what they think of doctors
- useful where research will lead to individual case studies rather than results which are applied to the whole population
17
Q
Sampling techniques - Self-selecting sample
A
- those involved volunteer to take part (given the choice to participate or decline) e.g. online survey
- can result in bias - someone who refuses to answer some of the questions
- not used in serious research
