Y1 Stats Flashcards
What are the advantages of a census?
Should give a completely accurate result
What are the disadvantages of a census?
Time consuming and expensive, hard to process large quantity of data.
What is the difference between a census and a sample?
Census measures every member of a population, sample is only from a subset of the population.
What are the advantages of a sample?
Less time consuming and expensive, less data to process, fewer have to respond.
What are the disadvantages of a sample?
Data may not be accurate, and there may not be enough data to give information about subgroups.
When can a census not be used?
If testing the product destroys it.
What are the advantages of simple random sampling?
Free of bias, easy and cheap for small populations, each sampling unit has a KNOWN chance of selection.
What are the disadvantages of simple random sampling?
Not suitable for a large population as it time consuming / disruptive, needs a sampling frame.
What are the advantages of systematic sampling?
Simple and quick, suitable for large populations.
What are the disadvantages of systematic sampling?
Need a sampling frame, if the frame isn’t random can introduce bias.
What are the advantages of stratified sampling?
Accurately reflects the population structure, guarantees proportional representation of groups.
What are the disadvantages of stratified sampling?
Population must be classified into ‘strata’, same as simple random sampling within each strata.
What are the advantages of quota sampling?
Allows a small group to represent population,
No sampling frame needed,
Quick, easy and cheap,
Allows for easy comparison between groups
What are the disadvantages of quota sampling?
Non-random sampling introduces bias,
Dividing population into groups can be inaccurate,
Hard to scale up,
Non-responses not recorded.
What are the advantages of opportunity sampling?
Easy, cheap
What are the disadvantages of opportunity sampling?
Unlikely to be representative, and accuracy depends a lot on researcher.
What is qualitative data, and what is quantitative data?
Qualitative = non-numerical observations
Quantitative = numerical observations
What is continuous data, and what is discrete data?
Continuous can take any value, discrete can only take specific values. Usually integers.
LDS - what is cloud cover measured in?
Oktas, eights of the sky covered by cloud
Where are the locations in the LDS?
UK: Camborne, Hurn, Heathrow, Leeming, Leuchars
Overseas: Beijing, Perth, Jacksonville
Rank the locations in the LDS from most to least far North.
Leuchars, Leeming, Heathrow, Hurn, Camborne, Beijing, Jacksonville, Perth
Which location in the LDS is in the southern hemisphere?
ONLY Perth
What are sampling frames?
Where individual units of a population are named or numbered.
How do you calculate the mean from a frequency table?
The sum of the products of data values and frequencies / sum of frequencies. Easier than it sounds.
How do you estimate the location of data through interpolation?
Assume data is evenly distributed within each class. So find width of class then how far along it you are.
How are variance and S.D related?
Standard deviation is the root of the variance.
How do you calculate variance?
Mean of the squares minus square of the mean
On the calc I use, where is the ‘frequency’ table?
2nd from left, assuming 1-var.
How does coding in the form y = (x - a )/ b work?
Mean is (mean of (x) - a )/b
S.D is original S.D / b
What do you do when finding quartiles?
Use n/4 or 3n/4. If not whole round up, if whole is halfway between that point and the one above it. How weird.
Where should the 2nd set of lines be on a box plot?
Either at highest/lowest non-anomalous value OR at the boundary for outliers.
What are cumulative freq. graphs useful for?
Finding the quartiles, median etc.
What can you comment on when comparing data?
Measure of location, measure of spread.
What are mutually exclusive events?
No overlap between events.
Events where P(A or B) = P(A) + P(B)
What are independent events?
Have no effect on each other.
P(A and B) = P(A) x P(B)
When can you model something using a binomial dist.?
Fixed number of trials, two possible outcomes, fixed success prob, trials are independent.
