ED 635 Action Research

Unit 8: Quantitative inferential

Unit overview

Descriptive statistics are designed to describe a distribution of numbers... while inferential statistics are used to determine if the characteristics of those numbers are statistically significantly different than a randomly selected group of numbers. You've seen language like this in research studies... p<.05 - well that is the result of an inferential statistic. What that means is English is that the probability of whatever (the mean difference between two groups of numbers... or the correlation coefficient between two sets of numbers...) is less than 5 chances out of 100. In education research we typically use .05 as the cut-off value for getting excited about things. Again, using other words... this means that the chances of getting this difference between means or this level of relatedness (in the correlation example) randomly are not very good (less than 5%). Or... the chances of these differences or relationships being related to something that actually happened in the real world is pretty good... at least 95% chance! See what I mean... give the reading a try before you squirm too much on this.

Moodle discussion

Use the Moodle forum to solicit help, ask questions, and wade through the t-test activity.

Unit content

Alright... after you feel like you have a general sense of descriptive statistics from unit 7... lets take a stab at inferential statistics here in unit 8.

First, download this (part 1 of chapter) and this (part 2 of chapter). Focus on pages 445-467 getting through the conversations about t-tests - both independent and non-independent samples t-tests!

Second, once you've give these readings the old college try (really, don't get frustrated here as this stuff is very difficult to be learning independently). Work on developing a rough understanding... let some of the details go. To help build an understanding of the basics, download this activity on t-tests and give it a whirl. Again, I will post an answer key in a few weeks. I lied again... here it is.

Third, down load this little summary that describes both internal and external threats to validity. Make sure you understand what threats to validity are... and how these are both similar to and different from threats to trustworthiness (which is how we talk about defensibility in the qualitative tradition).

Fourth, ready for something a little light-hearted that is still related to quantitative methods... try this!