statistics

census

measures every member of a population

(+) of census

accurate

(-) of census

expensice and testing may destroy

sampling units

individuals of a population

sampling frame

list of sampling units

simple random sampling

same chance of being selected

use random no. gen. or lottery sample

(+) simple random sampling

bias free

(-) simple random sampling

need sampling frame

systematic sampling

take every kth unit

k= (pop)/(sample size)

pick a random no. btw. 1 and k for starting no.

(+) systematic sampling

quick to use

(-) systematic sampling

sampling frame needed

stratified sampling

sample represents groups (strata) of a pop

(sample)/(pop size) x size of strata for each strata

pick randomly

(+) stratified sampling

reflects pop.

(-) stratified sampling

pop must be classified in strata

quota sampling

like stratified but strata filled by researcher/ interviewer

(+) quota sampling

no sampling frame

(-) quota sampling

non random, potential bias

opportunity sample

quota filled by those avaliable at the time

(+) opportunity sample

easy/ cheap

(-) opportunity sample

unlikely to be representative

qualitative data

non-numerical

quantitative data

numerical

either discrete or continuous

cities to know for weather stations

going from down up

Cambourne

Hurn

Heathrow

Leeming

Leuchars

uk stations which are winder and rainer

costal

uk weather stations that are warmer and have more sunlight

south

when large data set recorded

may-oct

for only 6 months

1987 and 2015

international stations

Perth, Australia

Beijing, China

Jacksonville, Florida, USA

international station where summer and winter switched around

Perth, Australia

international station where really hot and rainy in the summer and really cold in winter (extreme)

Beijing, China

international station where warm and hurricane prone

Jacksonville, Florida, USA

when were the hurricanes in Jacksonville, Florida, USA

oct 1987

oct 2015

tr data in large data set

trace

<0.05 mm rain

treat as 0 in all calc

n/a data in large data set

reading not available

can’t use in a sample

cloud cover data in large data set

oktas

discrete values

integers from 0-8

max gust data in large data set

knots

1 knot = 1.15 mph

when was great storm in uk

october 15th/ 16th 1987

(x bar)

(mean)

sum of x values of n

\overline{x} \frac{\sum x}{n}

(mean for grouped values)

\frac{\sum fx}{\sum f}

listed data and group data quartiles

LQ = \frac{n}{4}

Q2 =

UQ / Q3 =

if when find quartiles in listed data its a decimal

round up then find that position

if when find quartiles in listed data its a whole no.

find midpoint with next no.

percentiles in grouped data e.g 57th percentile or

P_{57}

0.57 x n

deciles in grouped data

10% chunks

D_3 = P_{30} = 0.3 x n

LINEAD INTERPOLATION for grouped data

ADD FLASHCARDS ON IT!!!

interquartile range

Q3 - Q1

(+) interquartile range

ignore extremes

interpercentile range eg. 10th to 90th

10th to 90th IPR = (make no. smaller and undeneath but)

P90 - P10

Variance \sigma^2=

Worded Variance

MSMSM

mean of the squares minus square of the means

Variance for grouped data

..

Remember \left(\sum fx^2\right)\ne\left(\sum fx\right)^2

when coding y = ax + b

whats y bar (put in symbols)

ybar = a xbar + b

so effected by both of those values

when coding y = ax + b

whats y standard deviation ( PUT SYMBOLS)

stardard deviation y = a standard deciation of x