Channel Title: Geoff Cumming Channel Description: The New Statistics Videos to support the understanding and use of The New Statistics Estimation is a much better way to do statistics than traditional null hypothesis significance testing (NHST). 'The New Statistics' (i.e., estimation, comprising effect sizes, confidence intervals, and meta-analysis) are not themselves new, but using them as the main strategy would be new for most researchers in many disciplines. And a wonderful step forward! The video 'Dance of the p values' gives just one reason why we should not use, or trust, p values. My book 'Understanding The New Statistics: Effect sizes, confidence intervals, and meta-analysis' (Routledge 2012) is intended to support anyone wanting to understand or use the new statistics. It is highly accessible, uses pictures and simulations to make statistical concepts intuitive, and has received wonderful reviews. More info at: www dot thenewstatistics dot com <video> 13.1 One proportion, for ITNS book Description: Frequencies, percentages, proportions. ESCI intro chapters 10-16, One proportion. The asymmetric CI on a proportion. This is a video to go with the book: Introduction to The New Statistics: Estimation, Open Science, and Beyond (Routledge 2017) by Geoff Cumming & Robert Calin-Jageman. See: www.thenewstatistics.com </video> <video> 14.1 One-way independent groups, for ITNS book Description: One-way independent groups design. Bushman example of advertising effectiveness. ESCI intro chapters 10-16, Ind groups comparisons. Planned comparisons, CI on a comparison. Rattan student motivation example. ESCI intro chapters 10-16, Ind groups contrasts. Planned contrasts of subset means, CI on a contrast. This is a video to go with the book: Introduction to The New Statistics: Estimation, Open Science, and Beyond (Routledge 2017) by Geoff Cumming & Robert Calin-Jageman. See: www.thenewstatistics.com </video> <video> 16.1 Future directions, for ITNS book Description: Open Science, the new statistics, the ten-step plan for research. Replicability of research in psychology, Many Labs 1, Reproducibility Project: Psychology. Preregistered review. Student participation in replication research. This is a video to go with the book: Introduction to The New Statistics: Estimation, Open Science, and Beyond (Routledge 2017) by Geoff Cumming & Robert Calin-Jageman. See: www.thenewstatistics.com </video> <video> 8.2 More paired design, for ITNS book Description: The paired design, Thomason 1 example. ESCI intro chapters 3-8, Data paired. Cohen’s d and dunbiased. Standardizer for the paired design. CI for delta. ESCI intro chapters 3-8, Summary paired. Thomason 2 example. No overlap rule for paired design. Comparing two designs. Carryover effects, counterbalancing, and parallel forms of a test. This is a video to go with the book: Introduction to The New Statistics: Estimation, Open Science, and Beyond (Routledge 2017) by Geoff Cumming & Robert Calin-Jageman. See: www.thenewstatistics.com </video> <video> 12.2 Regression, for ITNS book Description: Regression and inference. ESCI intro chapters 10-16, Scatterplots. Thomason 1 example. CI on slope, b. CI for mean Y at a particular X, curves for those CIs. Prediction interval for a single value of Y at a particular X, curves for those PIs. This is a video to go with the book: Introduction to The New Statistics: Estimation, Open Science, and Beyond (Routledge 2017) by Geoff Cumming & Robert Calin-Jageman. See: www.thenewstatistics.com </video> <video> 14.2 More one-way independent groups, for ITNS book Description: Halagappa Alzheimer’s example with mice. ESCI intro chapters 10-16, Ind groups contrasts. Planned contrasts of subset means, CI on a contrast. Planned and exploratory analysis, cherry picking. This is a video to go with the book: Introduction to The New Statistics: Estimation, Open Science, and Beyond (Routledge 2017) by Geoff Cumming & Robert Calin-Jageman. See: www.thenewstatistics.com </video> <video> 15.1 Two-way independent groups, for ITNS book Description: Two-way independent groups design. Frenda 2 × 2 false memory example. ESCI intro chapters 10-16, Ind groups 2 × 2. Main effects, simple main effects, with CIs. Interaction as difference of differences, with CI. This is a video to go with the book: Introduction to The New Statistics: Estimation, Open Science, and Beyond (Routledge 2017) by Geoff Cumming & Robert Calin-Jageman. See: www.thenewstatistics.com </video> <video> 12.1 Regression, for ITNS book Description: Linear regression of Y (predicted variable) on X (predictor variable). ESCI intro chapters 10-16, Scatterplots. Thomason 1 example. Residuals, equation of the line, intercept and slope. Standard scores, regression of ZY against ZX, with slope r. Regression for making predictions. This is a video to go with the book: Introduction to The New Statistics: Estimation, Open Science, and Beyond (Routledge 2017) by Geoff Cumming & Robert Calin-Jageman. See: www.thenewstatistics.com </video> <video> 14.3 One-way repeated measure, for ITNS book Description: Independent groups and repeated measure designs. Comparisons, contrasts, and their CIs. One-way repeated measure design. Donohue critical thinking example. ANOVA, an alternative approach to analyzing extended designs. Main effect of the one IV. This is a video to go with the book: Introduction to The New Statistics: Estimation, Open Science, and Beyond (Routledge 2017) by Geoff Cumming & Robert Calin-Jageman. See: www.thenewstatistics.com </video> <video> Significance Roulette 1 Description: The p value is extremely unreliable, enormously variable. In fact, running a typical experiment, to obtain a p value, is equivalent to spinning a roulette wheel marked with a very wide range of p values. We simply should not use p values. There are much better ways to draw conclusions from data. See www.thenewstatistics.com </video> <video> 15.2 More two-way designs, for ITNS book Description: Main effects and interaction. Rock/Classical, Party/Church example. Patterns of means, interaction as non-parallel lines. One IV as moderator of the effect of the other IV. RCT, two-way factorial design with one repeated measure, Hölzel meditation example. Two-way design with two repeated measures, Weisberg quality of explanation example. Two-way 3 × 2 mixed design, McDaniel study technique example. Two-way 3 × 5 mixed design, Chaix restricted feeding example with mice. General analysis strategy for extended designs. This is a video to go with the book: Introduction to The New Statistics: Estimation, Open Science, and Beyond (Routledge 2017) by Geoff Cumming & Robert Calin-Jageman. See: www.thenewstatistics.com </video> <video> Significance Roulette 2 Description: In Significance Roulette 1, I demonstrated that running a typical experiment is like spinning a roulette wheel to obtain a p value. Here I explain how running a replication of an initial experiment that obtained, for example, p = .05, is also like spinning a particular roulette wheel of p values. I then explain p intervals, the 80% prediction interval for the p value, following an initial experiment. These intervals are very long! The conclusion has to be that p values are enormously variable and totally unreliable. We simply should not use them. There are much better ways! See www.thenewstatistics.com </video> <video> 13.2 Two proportions, for ITNS book Description: The difference between two independent proportions. ESCI intro chapters 10-16, Two proportions. CI on the difference. Chi-square, an alternative approach to analyzing a 2 × 2 frequency table. Phi coefficient. This is a video to go with the book: Introduction to The New Statistics: Estimation, Open Science, and Beyond (Routledge 2017) by Geoff Cumming & Robert Calin-Jageman. See: www.thenewstatistics.com </video> <video> 16.2 More future directions, for ITNS book Description: Non-normal data, robust statistical techniques. Robust analysis for independent groups, trimming, and trimmed means. ESCI intro chapters 10-16, Robust two. N sex partners example. Archival and longitudinal data. Numerous DVs. Big data. Open Science and careful critical thought. This is a video to go with the book: Introduction to The New Statistics: Estimation, Open Science, and Beyond (Routledge 2017) by Geoff Cumming & Robert Calin-Jageman. See: www.thenewstatistics.com </video> <video> 10.1 Open Science, for ITNS book Description: Replication crisis, three Ioannidis problems, the p less than .05 imperative. Questionable research practices, p hacking. Psychological Science, Open Science policies, badges, the new statistics. Preregistration. Center for Open Science, Open Science Framework. This is a video to go with the book: Introduction to The New Statistics: Estimation, Open Science, and Beyond (Routledge 2017) by Geoff Cumming & Robert Calin-Jageman. See: www.thenewstatistics.com </video>

Description

The New Statistics
Videos to support the understanding and use of The New Statistics
Estimation is a much better way to do statistics than traditional null hypothesis significance testing (NHST). 'The New Statistics' (i.e., estimation, comprising effect sizes, confidence intervals, and meta-analysis) are not themselves new, but using them as the main strategy would be new for most researchers in many disciplines. And a wonderful step forward!
The video 'Dance of the p values' gives just one reason why we should not use, or trust, p values.
My book 'Understanding The New Statistics: Effect sizes, confidence intervals, and meta-analysis' (Routledge 2012) is intended to support anyone wanting to understand or use the new statistics. It is highly accessible, uses pictures and simulations to make statistical concepts intuitive, and has received wonderful reviews. More info at: www dot thenewstatistics dot com

Videos from channel