Post a composite review
Unpost a composite review
Search all MERLOT
Click here to go to your profile
Select to go to your workspace
Click here to go to your Dashboard Report
Click here to go to your Content Builder
Click here to log out
Search Terms
Enter username
Enter password
Please give at least one keyword of at least three characters for the search to work with. The more keywords you give, the better the search will work for you.
select OK to launch help window
cancel help

MERLOT II


    

Peer Review


Comparison of Cultures: A t-test Tutorial

by Dale Berger
 

Ratings

Overall Rating:

3.5 stars
Content Quality: 3.25 stars
Effectiveness: 3.25 stars
Ease of Use: 3.75 stars
Reviewed: Jan 19, 2010 by Statistics Editorial Board
Overview: The following scenario is posited: Imagine you are a researcher interested in cross-cultural differences in child rearing. You have data on the distance measured between mother and child for two different countries: Belize and Samoa. You have data for a sample of 4-year old children from each country regarding how far each child is, on average, from his or her mother at a predetermined observation time. How would you proceed? The user is then asked a series of branching-type questions that includes stating null and alternative hypotheses, running an independent t-test, and interpreting his/her result.
Learning Goals: This is intended to communicate all of the steps of a statistical analysis of two groups, from the construction of the data, a preliminary data analysis, setting up the hypotheses, computing the test statistic, and drawing the conclusions.
Target Student Population: Any statistics class that is introducing hypothesis testing.
Prerequisite Knowledge or Skills: Should already have some familiarity with hypothesis testing and the independent t-test.
Type of Material: Tutorial, a set of linked web pages with graphics, tabular content, and text explanations of concepts.
Recommended Uses: Could be used after students have learned the basics of hypothesis testing and the t-test (especially 2-sample comparison of means) to help them make decisions on when to use which statistical test.
Technical Requirements: Any internet browser program can be used for this tutorial.

Evaluation and Observation

Content Quality

Rating: 3.25 stars
Strengths: Goals are stated within different parts of the module. There is an "Ask the expert" link, which will lead the student through a series of scaffolding questions to (hopefully) arrive at the correct answer. Follow-up questions are provided at the conclusion of the lesson that encourages students to interpret what they just learned.
Concerns: There are some dead-ends; that is, in some parts of the module, the user must use the back button several times to get back to the beginning. Some of the discussion on the web pages could be improved.

Potential Effectiveness as a Teaching Tool

Rating: 3.25 stars
Strengths: Has effective explanations of correct answers as well as good explanations when user selects the incorrect path. Exposure to mini case studies like this are helpful, since they communicate the entire process of a statistical investigation.
Concerns: Should be updated to include p-value method for determining significance. Degrees of freedom explanation does not necessarily match that in some major textbooks. User may get frustrated with dead links and dead-end paths. Should link to one of their own applets to let user put in mean, sd to calculate the mean. (http://wise.cgu.edu/hypothesis/hypothesis_applet.asp)

Ease of Use for Both Students and Faculty

Rating: 3.75 stars
Strengths: It was easy to navigate through the web pages and respond to the questions posed.
Concerns: Students might get confused by clicking on a wrong link. It could be improved by a table of contents at top that gives links to all pages.

Other Issues and Comments: I thought the discussion about a one or two-tailed test could be improved. The bell-shaped curve was not explained and the labels on the x-axis were confusing. The explanation on why other analyzes, like a correlation analysis, were weak.