Sampling Distribution

[mnp2@pdx.edu]

I’m trying to figure out if I can use CODAP for teaching sampling distributions, or just for data visualization/exploration. I have two questions:

First, is there a way to adjust the percentages in the spinner? All I can find is to set them in the mixer, and then switch to the spinner, but that’s not something I can expect students to do.

Second, is there a tool like TinkerPlots’ “Collect Statistics” tool? It looks like I’m able to collect multiple samples at once. But the most natural way for students to display these samples is as a single large sample(!!!). And even if I make “sample” the y-axis I can’t see a sampling distribution, only the distributions of lots of individual samples. I can see how that would be helpful, but not if that’s all the further I can go. (Perhaps I could program another column, but because the data is all displayed in a single column, it seems like that would be difficult for students.)

[Kathy Shafer]

This is a great question. In the file linked below, I illustrate the Central Limit Theorem. This involved drawing samples of a given size, finding the mean for each sample, followed by creating the distributions of the means for the samples. Since I used canned data, I already know the population mean, but you can find that by sampling, without replacement, using the n value for your population. It looks like I ran the sampler twice for n=75, but you get the idea.

https://codap.concord.org/app/#shared=130461

Note that the sampler appears to have lost the collection of 200 salary’s. I have found the program to be a little buggy in these types of problems and I don’t have time to play with it this morning.

If you have any questions, I have other sample files available. Also, please let me know if this is not helpful.

Kathy

[trharri2@ncsu.edu]

When you use the sampler, it creates a hierarchical table of experiments/samples/items.  To create a sampling distribution, the key is to create statistics at the sample level (click the plus sign next to the “sample” section of the table).  For example, here is a file where I created a statistic for the sample proportion (“proportionA”) and graphed it.  You can also use the sampler to collect from a population to do the distribution of sample means like here.

You could also extend it to something like this where I’ve run the sampler three different times with different sample sizes, and put the sample_size attribute on the y-axis to show the effect of sample size on the sample distribution.

[Hollylynne Lee]

The above examples are likely helping you imagine how you might build something to help develop better understandings of a sampling distribution.

 

As aprt of the ESTEEM project, we have two resources that may be of further help.

1. a brief video that explains some basics about using the sampler and woring with heirarchical tables.

2. a task about investigating the typical CO2 emissions in samples of vehicles.

[mnp2@pdx.edu]

That’s extremely helpful. Thanks!

[Anonymous]

Thanks for sharing answers. I have same problem

[Anonymous]

It’s really helpful, thanks

[Author]

Posts