Understanding The New Statistics
Effect Sizes, Confidence Intervals, and Meta-Analysis
Geoff Cumming
- 536 pagine
- English
- ePUB (disponibile sull'app)
- Disponibile su iOS e Android
Understanding The New Statistics
Effect Sizes, Confidence Intervals, and Meta-Analysis
Geoff Cumming
Informazioni sul libro
This is the first book to introduce the new statistics - effect sizes, confidence intervals, and meta-analysis - in an accessible way. It is chock full of practical examples and tips on how to analyze and report research results using these techniques. The book is invaluable to readers interested in meeting the new APA Publication Manual guidelines by adopting the new statistics - which are more informative than null hypothesis significance testing, and becoming widely used in many disciplines.
Accompanying the book is the Exploratory Software for Confidence Intervals (ESCI) package, free software that runs under Excel and is accessible at www.thenewstatistics.com. The book's exercises use ESCI's simulations, which are highly visual and interactive, to engage users and encourage exploration. Working with the simulations strengthens understanding of key statistical ideas. There are also many examples, and detailed guidance to show readers how to analyze their own data using the new statistics, and practical strategies for interpreting the results. A particular strength of the book is its explanation of meta-analysis, using simple diagrams and examples. Understanding meta-analysis is increasingly important, even at undergraduate levels, because medicine, psychology and many other disciplines now use meta-analysis to assemble the evidence needed for evidence-based practice.
The book's pedagogical program, built on cognitive science principles, reinforces learning:
-
- Boxes provide "evidence-based" advice on the most effective statistical techniques.
- Numerous examples reinforce learning, and show that many disciplines are using the new statistics.
- Graphs are tied in with ESCI to make important concepts vividly clear and memorable.
- Opening overviews and end of chapter take-home messages summarize key points.
- Exercises encourage exploration, deep understanding, and practical applications.
This highly accessible book is intended as the core text for any course that emphasizes the new statistics, or as a supplementary text for graduate and/or advanced undergraduate courses in statistics and research methods in departments of psychology, education, human development, nursing, and natural, social, and life sciences. Researchers and practitioners interested in understanding the new statistics, and future published research, will also appreciate this book. A basic familiarity with introductory statistics is assumed.
Domande frequenti
Informazioni
1
Introduction to The New Statistics
- Null hypothesis significance testing (NHST) and confidence intervals (CIs) as two different ways to present research results
- Meta-analysis as a way to combine results, and thus a third way to present them
- The desirability of shifting emphasis from NHST to CIs and meta-analysis, which are important parts of what I'm calling the new statistics
- Three ways of thinking that correspond to the three ways to present results
- Evidence-based practice in statistics, statistical cognition, and some relevant evidence
- ESCI (pronounced “ESS-key,” Exploratory Software for Confidence Intervals)
A Familiar Situation: Lucky–Noluck
First Presentation: NHST
Only two studies have evaluated the therapeutic effectiveness of a new treatment for insomnia. Lucky (2008) used two independent groups each of size N = 22, and Noluck (2008) used two groups each with N = 18. Each study reported the difference between the means for the new treatment and the current treatment.Lucky (2008) found that the new treatment showed a statistically significant advantage over the current treatment: M (difference) = 3.61, SD (difference) = 6.97, t(42) = 2.43, p = .02. The study by Noluck (2008) found no statistically significant difference between the two treatment means: M(difference) = 2.23, SD(difference) = 7.59, t(34) = 1.25, p = .22.
- Inconsistent “The Lucky result is clearly statistically significant at the .05 level, whereas the Noluck result is clearly not statistically significant. The two results conflict. We can't say whether the treatment is effective, and we should examine the two studies to try to find out why one found an effect and the other didn't. Further research is required to investigate why the treatment works in some cases, but not others.”
- Equivocal “One result is statistically significant, and the other statistically nonsignificant, although the two are in the same direction. We have equivocal findings and can't say whether the treatment is effective. Further research is required.”
- Consistent “The two results are in the same direction, and the size of the mean difference is fairly similar in the two studies. The two studies therefore reinforce each other, even though one is statistically significant and the other is not. The two results are consistent and, considered together, provide fairly strong evidence that the treatment is effective.”
Second Presentation: Confidence Intervals
Only two studies have evaluated the therapeutic effectiveness of a new treatment for insomnia. Lucky (2008) used two independent groups each of size N = 22, and Noluck (2008) used two groups each with N = 18. Figure 1.1 reports for each study the difference between the means for the new treatment and the current treatment, with the 95% confidence interval on that difference.
Third Presentation: Meta-Analysis
Only two studies have evaluated the therapeutic effectiveness of a new treatment for insomnia. Lucky (2008) used two independent groups each of size N = 22, and Noluck (2008) used two groups each with N = 18. Each study reported the difference between the means for the new treatment and the current treatment.