Stats Flashcards
(22 cards)
Census (+&-)
Measures every member of a population
+ accurate result
- expensive/ testing may destroy
Sampling units
Individuals of population
Sampling frame
List of sampling units
Simple random sampling (+&-)
Same chance -> random no. Or lottery
+unbiased
- need sampling frame
Systemic sampling (+&-)
Take kth unit, k=pop/sample, random no. & that’s starting point
+ quick to use
- need sampling frame
Stratified sampling (+&-)
Sample represents strata of a pop, sample/pop X strata & pick random
+ reflects pop
- pop must be in strata
Quota sampling NR (+&-)
Like Stratford but strata filled by interviewer
+ no sampling frame
- non random, potential bias
Opportunity sampling NR (+&-)
Quota filled by available
+ easy and cheap
- biased & not representative
Name all UK large data set
(North to south)
Leuchars
Leaning
Heathrow
Hurn
Cambourne
International large data set places
Perth, Australia (swapped weather)
Beijing, China - hot & rainy& cold- seasonal
Jacksonville, Florida USA (hurricanes in oct 87&15)
Cloud cover
Oktas, discrete values 0-8
Max gust
Knots, 1kn=1.15mph, great storm in uk oct’87
Means and quartiles for listed data
Mean- leave
Quartiles-
If decimal, round up
If whole, find midpoint with next one
Grouped data mean and quartiles
No rounding, use linear interpolation
Coding for y=ax+b (mean & standard dev.)
Ymean = axmean +b (affected by all)
Standard deviation y= ax standard dev (only affected by multiplication)
Mean for frequency data
Sum of fXx/ sum of frequency
Standard dev for frequency data
Root: sum of f X x^2 / sum of f. - (mean)^2
Area of histogram bar
Freq X k
Daily mean visibility measurement
Dm (decametres)
Daily mean pressure measurement
hPa (hectopascals)
Conditional probability formula
P(A|B) = P(AnB) / P(B)
Formula for standard normal distribution
Z = y- mean / standard deviation
