Statistics for HCI. Alan Dix

Скачать книгу

       PART II Doing It –If not p then What

       5 Probing the unknown

       5.1 Recall the job of statistics

       5.2 Conditional probability

       5.3 Likelihood

       5.4 Statistical reasoning

       5.5 Types of statistics

       6 Traditional statistics

       6.1 Hypothesis testing

       6.1.1 The significance level –5 percent and all that

       6.1.2 But what does it mean?

       6.1.3 Non-significant results

       6.1.4 In summary—significance

       6.2 Confidence intervals

       6.2.1 The interval

       6.2.2 Important as well as significant?

       6.2.3 Don’t forget

       7 Bayesian methods

       7.1 Detecting the Martian invasion

       7.2 Quantifying prior belief

       7.3 Bayes for intelligent interfaces

       7.3.1 Bayes as a statistical method

       7.4 How do you get the prior?

       7.5 Handling multiple evidence

       7.6 Internecine warfare

       8 Common issues

       8.1 Cherry picking

       8.1.1 Multiple tests

       8.1.2 Multiple statistics

       8.1.3 Outliers

       8.1.4 Post-hoc hypothesis

       8.1.5 The file drawer effect

       8.2 Inter-related factors

       8.2.1 Non-independently controllable factors

       8.2.2 Correlated features

       8.3 Everything is random

       8.4 The same or worse

       8.4.1 Everything is unlikely

       8.4.2 Numeric data

       8.4.3 More complex ‘or worse’

       8.4.4 Post-hoc corrections

       8.5 Simulation and empirical methods

       8.6 What you can say—phenomena and statisticians

       9 Differences and distinctions

       9.1 Philosophical differences

       9.1.1 What do we know about the world?

       9.1.2 Not so different

       9.2 So which is it?

       9.2.1 The statistical crisis

       9.2.2 Alternative statistics

       9.3 On balance (my advice)

       9.4 For both

       9.5 Endnote

       PART III Design and Interpretation

       10 Gaining power –the dreaded ‘too few participants’

Скачать книгу