types of data

• 4 levels of data measurement


  1. nominal (simplest data)
  2. ordinal (more precise than nominal
  3. & 4. interval and ratio (most precise)
• data can be treated as nominal

  • results are in totals in two or more named categories
  • this shows the number of times something occured
  • there are no scores for individual participants
  • likely to be collected from closed questions in self reports or from strucutred observations
  • for example: the number of people who helped or didnt help
• data can be treated as ordinal

  • results are able to be placed in rank order (e.g. in positions such as 1st, 2nd and 3rd)
  • the difference between each rating, rank or score is not known as it is a non-scientific scale for example, coming 1st, 2nd, 3rd in a beauty contest
  • does not have to be equal for example, putting class test scores in rank order does not have to be equal 1st place could be 90/100 2nd 88/100 but 3rd 80/100
  • each participant has an individual score
• data can be treated as interval/ratio

  • results are made up of numbers that come from a scientific scale of equal or known units
  • each participant has an individual score
  • interval data can go into negative values for example, temperature
  • ratio data has an absolute zero and so are often mathematical units for example, weight, distance or time
  • the term ratio data is not mentioned on the spec so we will assume that both types are covered under the term interval data
• • we can treat interval/ratio data as it is or treat it as ordinal or nominal(convert it)
  • ordinal can be treated as it is but can also be treated (converted) as nominal
  • nominal cannot be made more precise & so can only be treated this way
• nominal data


  strength:

  easy to analyse

  weakness:

  does not allow for comparisons between participants
• ordinal data


  strengths:

  allows for some comparison between participants, we can put participants in order from first to last

  the data can be simplifies and treated as nominal data

  weaknesses:

  the date does not always allow us to see the difference between participants scores, only that someone has score better/worse than someone else

  the data is measured in non-scientific units, such as strengths of beliefs, scores on a test etc.
• interval data(and ratio data)

  strengths:

  provides measurement of participants in universally accepted units. this allows for detailed comparison between participants including from first to last and the difference between partcipants.

  data can be simplifies and treated as ordinal and nominal

  no weaknesses
• mean


  most suitable to use for data treated as interval/ratio but can also be used for date treated as ordinal

  :) uses all of the raw data

  :( is affected by extreme scores so may be misleading
• median


  most suitable for data treated as ordinal & can also be used for data treated as interval/ratio

  :)not affected by extreme scores

  :(can be distorted by small samples
• mode


  2 modes- bimodal

  more than 2- multi modal

  most suitable for data treated as nominal

  :)not influenced by extreme scores and can show the most popular value

  :( doesnt use all the date so may not be repesentative
• range

  highest - lowest +1 (for measurement error)


  suitable for data treated as interval/ratio & ordinal

  :) easy to calculate

  :( can be influenced by extreme scores and so it may be misleading as it tells us nothing about the distribution of other scores
• variance- tells us more baout the range. rather than looking at the extremes of the data set, the variance considers the difference between each data point and the mean, this is called the deviation


  :) takes every score into account and is therefore not affected by outliers as much as the range

  :( harder to caluclate the range + scores not in line
• standard deviation- spread of data around the mean

  helps us understand whether data is closely clustered around the mean or very spread out.

  shares the same strengths and weaknesses of the variance in that it uses all of the data in the data set, but requires a fairly complex calculation with the range.

  :) more precise measurement

  :( harder to calculate
• standard deviation formula