Sin título

Data Analysis

1. Qualitative

   • Converts raw data into a final narrative using a step-by-step process
     • The narrative summarises in a prosaic form
     • Analytical process
       • Manual analysis
       • Analysis using software
         • May gain extra credit
         • use of software not a substitute for understanding
   • 10 steps
     • (1) The research questions and objectives
     • (2) Interview questions and prompts.
     • (3) Verbatim transcripts of interviews first copy; separate files for each interview, saved as file 3a, 3b, etc.
     • (4) Verbatim transcripts of interviews second copy; originating from file 3.
     • (5) Content analysis; frequency counts.
     • (6) A basket of labels.
     • (7) Verbatim transcripts of interviews third copy; originating from file 4.
     • (8) Verbatim transcripts of interviews fourth copy
     • (9) A new file, comprising tables.
     • (10) The final narrative.

2. Quantitative

   • Descriptive
     • Statistical analysis
       • Manually
         • Can produce a good dissertation
       • Use of software
         • Statistical Package for the Social Sciences
         • Minitab
         • Powerful tools
           • Far better than Excell
       • Types of tests
         • Cronbach's alpha
         • Kendal's tau
         • Factor analysis
         • Gain extra mark if understanding of tests is shown
         • None-parametric
           • Pearson's
           • Chi-square
           • Mann-Whitney
           • Wilcoxon
           • Spearman's rho
           • Can not be done in excell
     • May need to rank data
       • Especially if you have amount of data
     • Normal distribution
       • For very large samples
       • Mean, median and mode are measures
     • Measures of spread
       • Standard deviation
       • Range
       • Variance
   • Inferential
     • Measurements
       • Categorical
       • Ordinal
       • Interval
       • Ratio
       • Level of measurements dictates the tests that are used
         • Chi-square
           • It compares the actual distribution of a set of frequency counts with theoretical or expected frequency counts.
             • If the actual frequencies are different to the expected frequencies, it is inferred that the difference is due to an association between the variables;
               • If so null hypothesis is rejected
               • Proves a relationship between the IV and DV
           • IV and DV measured at categorical/ nominal level
           • Larger sample size give more reliable data
         • Difference in means
           • Richer than Chi-square
             • Do more than compare differences means; they compare the distribution of numbers in a data set.
           • Mean scores are a key output
           • One variable at the categorical/ nominal level
             • Other variable ordinal or interval
           • Classify your data
             • Unrelated and non-parametric .
             • Related and non-parametric .
             • Unrelated and parametric .
             • Related and parametric
           • Types of tests
             • Mann-Whitney
             • Parametric
             • Wilcoxcon
         • Correlations
           • Richest
           • Correlation coefficients are key output
           • Both variables need to be at the ordinal, interval or ratio level
           • Both data sets parametric
             • Pearson correlation
           • One or both non-parametric
             • Spearman Rho test
               • Calculate manually or with the use of software
     • Independent and dependant variables
     • Consider probability values
       • Ideally 0.05