Hello Keith,

Geoff Cumming of La Trobe University has put together a nice collection of videos on a variety of introductory topics. 

http://www.latrobe.edu.au/psy/research/cognitive-and-developmental-psychology/esci

They can also be accessed through I-Tunes. 

Branko

I like a coursera course by Brian Caffo. He also uploaded his coursera lecture on Youtube: https://www.youtube.com/playlist?list=PLpl-gQkQivXiBmGyzLrUjzsblmQsLtkzJ. 

He also wrote a little book about statistical inference: https://leanpub.com/LittleInferenceBook

Cost is often an issue, so learn R - it is free off the CRAN web site.

There is the underlying distribution. For fun, use the Normal distribution. In R you can get as many samples from the Normal distribution as you want. The function is rnorm(X, mean=0, sd=1) for the Normal distribution with mean of zero and standard deviation of 1. X is the number of samples. Now make a histogram.

p

Keith -

For articles, I hope you will include something on the chronic problem of misinterpreting p-values without a power analysis or other sensitivity analysis. (Note first attachment.)

As for sample sizes, I see one thing in common to continuous data problems, and that is a need for a preliminary estimate of sigma. (Second attachment notes comparison between simple random

sampling 'formula' in Cochran compared to a very different type of sampling.)

Cheers - Jim

Article Practical Interpretation of Hypothesis Tests - letter to the...

Conference Paper Projected Variance for the Model-based Classical Ratio Estim...

Please refer to Sample Survey Resources Server available at  http://sample.iasri.res.in/ssrs/

Can you use a simple article or example of stereotypes, racism sexism or if you're really brave you can do mental health (statistics regarding mental health are widely available on the net i.e.. google depression in college students)? That will hit students hard while teaching them about stigma as well as statistics. I have a couple of PDF's on hand if needed.

WOW! thanks for the suggestion.. I will share this with the research group.

I will need more lit review to justify this... Thanks for your help.

Andrew, Could you share your PDF's teaching tool?

Keith

I work with Keith and in addition to the question he asked above... We are also interested in any research that talks about what the most common "misunderstood" concepts are?  What are some of the core concepts that students need to come into Statistics with?  If students struggle with Statistics, what is the best path toward closing their knowledge gaps? If anyone has any research or possible resources related to underserved populations and/or marginalized groups related to learning statistics. It would be most appreciated!  Thanks!

Gary -

You said "We are also interested in any research that talks about what the most common 'misunderstood' concepts are?"

That would likely be the p-value as I noted above. There are many comments on that on ResearchGate threads. (When possible I'd replace with confidence intervals.).

Another problem would be R-square values which are also much misused and also not very useful. (Estimated variance of the prediction error can be helpful.)

Cheers - Jim

Failure to reject the null hypothesis is not equivalent to proving that the null hypothesis is true. I ran into this recently where there was a debate on whether to keep a data point. We both agreed that the data point was flawed. However, by keeping the data point they could reject the null hypothesis and that outcome made biological sense. If the data point was rejected then the null hypothesis was not rejected and that was not consistent with the biological interpretation. The problem is that the argument only works if one has accepted the null hypothesis because one failed to reject -- when the right answer is that the outcome is inconclusive.

A long time ago I also saw this: Many statistical tests assume that the data are normally distributed. There are tests to see if data are normally distributed. However, justifying the use of parametric tests because you failed to reject the null hypothesis that the data are normally distributed doesn't work.

Keith.

I don't understand what you mean PDF teaching tool. Could you kindly explain to my simplistic mind?

Andrew -

Thank you for your input... We are following up on those suggestions... 

While we are engaged in creating and disseminating the preliminary versions of these resources, we do need to find relevant literature to justify the money and time needed for this project. We are currently in a "proof of concept" and testing of materials phase. When it comes to writing this up the relevant literature will be of use then as well. Any resources will be welcomed.

Gary

Andrew,

You mentioned: That will hit students hard while teaching them about stigma as well as statistics. I have a couple of PDF's on hand if needed.

I would like to see some of your PDF's and see what you have to offer. Sorry that I was not that clear..

Keith

Oh don't worry about it, I just didn't quite understand what was meant by teaching tool.

On the other hand I would use psychology to explain phenomena and incorporate statistics. Like how personality is along a normal curve and when you get the highs and lows of a trait (for example impulsiveness) then you typically have a disorder present. Herd behavior is basically stereotypes and how following the crowd can empower or hurt you, depending how you look at it.

The paper on norms is PERFECT to explain statistics. 

But if I were you or working with you I would gather information on various student based groups (like sexism, racism and all that jazz) and use their teachings and plus into a statistical framework to explain the methods.

If you require more help, please feel free to message me.