Chapter 1
Simple random sampling - every sample has equal chance.
No bias, easy, cheap, each s.unit known & equal chance of selection
Not suitable large (time consuming, disruptive, expensive), need sampling frame
Systematic sampling - unit chosen at regular intervals from ordered list.
Simple, quick, suitable for large sample/populations
Need sampling frame, can introduce bias
Stratified sampling - population divided into exclusive strata & sample from each taken.
Accurately reflects population structure, guarantee proportional result
Population must be spilt into distinct strata, each strata - same dis. as simple
Quota sampling - researcher selects sample that reflects population characteristics
Small & representative sample, no sampling frame, quick, easy, inexpensive, allows easy
comparison between different groups
Non-random = bias, grouping (costly/inaccurate), bigger scope = more groups non-responses
not recorded
Opportunity sampling - taken from people available at time of study + fit criteria
Easy, cheap
Highly dependant on researcher, unlikely to be representative
Data types;
Quantitative = numerical
Qualitative = non-numerical
Discreet = specific values/whole numbers
Continuous = any range (decimals)
In a table; - Class boundaries = min. & max. values
- Midpoint = average of class boundaries
- Class width = diff. between upper and lower class boundaries
UK : Leuchars, Leeming, Heathrow, Hern, Cambourne
Global : Beijing, Jacksonville, Perth
Daily mean temp - °C
Daily total rainfall - mm
Daily mean pressure - hectopascals (hPa)
Daily mean wind speed & direction - midnight-midnight, Beaufort scale
, (UK recorded only;)
Daily total sunshine - nearest 1/10th of an hour
Daily max. gust - knot (1 kn ~ 1.15 mph)
Daily max. relative humidity - %
Daily mean cloud cover - oktas (1/8ths)
Daily mean visibility - decameters (Dm)
Chapter 2
Measure of location - single value, describes position in data set
Measure of central tendency - single value in absolute centre
Mode/modal - most frequently occurring (for any type, not helpful is each value occurs once)
Median - middle value when put in order (quantitative data - not affected by extreme values)
(quantitative, uses all values = true measure of data, but is affected by extremes)
Lower quartile - ¼ through data set (Q1)
Median - Q2
Upper quartile - ¾ through data set (Q3)
Percentiles - split data into 100 - i.e, 10th percentile = 1/10 into data set (P10)
Range - difference between largest and smallest values
Interquartile range (IQR) - difference between UQ and LQ (Q3 - Q1)
Interpercentile range - varies based of percentiles used
