Data recording, analysis and presentation

• nominal level data

  this is the lowest level of data. it is just a headcount of the number of participants who do one thing as opposed to another

  e.g. number of males vs females wearing red in the cafeteria
• ordinal level data


  this is the medium level of data. the measures used are not carefully calibrated, so no attempt is made to analyse how much further highest is from second highest and second highest is from third highest. analysis is made of individual scores achieved by participants, but only in relation to each other

  e.g. students are asked to rate their enjoyment of apples on a scale of 1-10
• interval or ratio level data


  this is the highest level of data. it involves the use of carefully calibrated instruments of measurement so the sizes of the gaps between the highest score and second highest score can be measured. analysis is made of the actual scores achieved by individual participants.

  e.g. individuals are timed on how long it takes them to eat an apple. this is measured in seconds and minutes
• calculating the mode


  the most common value
• calculating the median


  the middle value, or the point between the two middle values
• calculating the range


  the difference between the largest value and the smallest value
• calculating the mean


  the result of all of the values added together divided by the amount of values
• calculating the variance


  -calculate the mean score of each condition in the experiment

  -for each participant, you then subtract the mean score from their score

  -square each difference

  -add all of the differences together

  -divide the total of the differences by n-1 (the number of participants minus 1)
• standard deviation

  the square root of the variance
• disadvantages of variance

  because it is the average of the squared differences from the mean, the figure that is calculated is not being expressed in the same units that the data in the original data set
• advantages of the variance

  it gives us an idea of how spread out the scores are from the mean
• advantages of standard deviation

  it is more useful than the variance as you can refer to the standard deviation in terms of the original category, it can tell us much more detail about how spread out the data is around the mean, median and mode
• disadvantages of standard deviation

  it can be time consuming to calculate and does not give you the full range of the data
• for spearman's rho and chi squares

  the calculated value must be higher than the critical value in order to accept the null hypothesis and reject the alternative hypothesis
• for mann whitney, binomial sign and wilcoxon signed ranks


  the calculated value must be lower than the critical value in order to accept the alternative hypothesis and reject the null hypothesis
• type 1 error


  accepting the alternative hypothesis when you should reject it (false positive)
• type 2 error


  rejecting the alternative hypothesis when you should accept (false negative)
• parametric tests


  T-tests and Pearson's product moment

  you use these when you have interval level data, there is a normal distribution of the results and all groups in the research have similar variances
• Binomial sign test


  for nominal data and repeated measures design

  1. indicate the direction of difference for changes in attitude (if the participant has a more favourable attitude put a + and for less favourable put -)
  2. count the number of times the least frequent sign (+/-) appears and call this your calculated value of X
  3. compare your calculated value to the critical value which you find in the table
  4. if the calculated value is lower than the critical value, reject the alternative hypothesis and accept the null hypothesis
• Chi square (independent/nominal)

  1.calculate the expected frequency of each cell (number of rows x number of columns divided by the total number of results)

  2. calculate the difference of the observed frequency and expected frequency and square each difference

  3. divide each squared difference by the observed frequency of that cell and add up all the values to get the calculated value

  4. calculate the degrees of freedom (number of rows-1) x (number of columns-1) to find the critical value

  5. if your calculated value is equal or higher to the critical value, accept the alternative hypothesis
• Wilcoxon signed ranks (repeated/ordinal)

  1. ignore the scores of any participant who scored the same in both conditions
  2. calculate the difference between each score in condition A and its equivalent in condition B. subtract in the same direction each time
  3. rank the differences, ignoring the sign.
  4. add together the ranks for the least frequent sign, and call this your calculated value of T
  5. find the critical value, accept the alternative hypothesis if the calculated value is lower than the critical value
• Mann-Whitney U-test (independent/ordinal)


  1. places the response to the critical question in rank order, rank them all together in one table
  2. for the first condition, work out the number of participants multiplied by the number of participants +1 and divide this by 2, then subtract this sum from the total sum of the ranks for the scores, this is your U value
  3. repeat for the second condition
  4. pick the lowest U value and this is your calculated value, if this is lower than the critical value, accept the alternative hypothesis
• Spearman's Rho (correlation/ordinal)


  1. rank the scores for the fist co variable, enter these rank values into the third column of the table
  2. rank the scores for the second co-variable, enter these rank values into the fourth column of the table
  3. calculate the difference between the rank value of the first co-variable and the rank value of the second co-variable. enter this difference into the fifth column
  4. square each difference and place in the sixth column
  5. add up all of the difference values
  6. apply this to the equation
  7. 1 – 6(∑d2)/N(N*2-1)
• representativeness refers to the sample used in research 


  if the sample is diverse and includes people from different ages, genders, occupations, education levels, etc, it will be more representative of the target population
• generalisability 


  refers to the results of the research. if the sample used in the research is biased and not very diverse, the results cannot be generalised to everyone in the target population
• a statement of statistical significance

  include:

  1. the test used
  2. the observed and critical value(s)
  3. the direction of the hypotheses
  4. the chosen value of p (e.g. P<0.05)
  5. the degrees of freedom (if necessary)
  6. the conclusion, in terms of the null hypothesis