Statistics

• Assumptions for hypothesis testing?

  1. fixed probability
  2. fixed no. Trials
  3. 2 possible outcomes
  4. independent of eachother
• If calculated value is less than sig level, reject H0
• Percentile px = x(n+1) / 100 where n is population total and x is wanted percentile
• Explanatory variable is the independent variable
• response variable is the dependant variable
• what is a census?

  when you measure every member of the population
• What is a sample?

  when you select a certain amount of people from a subset of a population
• Sampling units are the individuals of a population
• Sampling frame is a numbered list of sampling units
• Simple random sampling requires a sampling frame and is when you use a computer to generate random numbers until filled sample.
• Systematic sampling is when you take every certain number from sampling frame. population / sample size gives the every nth person you should choose. Random number generator for first person in 1-n. add n each time.
• Stratified is when the population is divided into mutually exclusive data. accurate proportions. N sampled in stratum = no. In stratum/ population. X sample size. Then SRS in stratum
• Quota sampling reflects characteristics of a whole pop. Divide sample into groups by characteristics, the proportion of group determines the size of characteristics out of whole population. Fill all quotas. bias
• Opportunity or convenience sampling is when you sample people there at the time. when it’s convenient to you.
• Quantitive data is numerical observations
• Qualitive data is non numerical
• Uk places in large data set include?

  Leuchars (Scotland), Leeming (central), Heathrow, Hurn (Hampshire) and cambone (SW coast)
• World wide locations in large data set?

  Jacksonville, USA. Beijing, China . Perth, Australia.
• Daily mean pressure is measured in hPa
• daily mean wind speed is measured in knots (1 Kn = 1.15mph)
• Daily max humidity is given in %. if > 95% then it is misty
• Approximating the binomial using normal is only valid when n is large and p close to 0.5 (as normal is symmetrical)
• When approx binomial with normal, standard deviation = root/ np (1-p) and mean = np
• When approximating the binomial using normal you must apply continuity correction
• Z = x - mean / standard deviation
• When hypothesis testing for the mean, standard deviation = root /( standard deviation squared / n )
• apply continutity correction to : p(x<9)

  p(Y<8.5)
• Apply continuity correction to: p(x>5)

  p( Y>5.5)
• Apply continuity correction to : p(x=9)

  p(8.5<Y<9.5)
• Apply continuity correction to p(x</= 3)

  P(Y<3.5)
• Why do you apply continuity correction?


  As binomial is discrete and normal is continuous.
• What is p(x>/= X) the same as?

  1 - p(x </= (X-1))
• What is typed when rainfall is less than 0.05mm?

  trace Tr
• What is p(X < N) the same as?

  P(X</= (N-1) )
• Large data set is from May - October
• daily wind speed is in knots
• Daily max humidity in %
• Daily mean visibility in Dm
• Daily mean cloud cover in 1/8s
• Daily mean cloud cover on scale of 0-8