11.1 Development

Topic:

Is There Cognitive Decline with Age Regardless of Stimulation?

Statistical Concepts Covered:

In this applet, you’ll expand on your knowledge of correlation, and be introduced to the concept of z scores and standardization.

Introduction

Development happens over the entire life span, beginning from conception and continuing until death. As we age, things change, including our physical appearance, attitudes, interests, and emotions. Even though we tend to think of these changes as declines, there are some areas in which we continue to excel or increase in over time. The Development chapter covers how our semantic and episodic memory change over time, with the former increasing or staying the same and the latter decreasing as we age. Research conducted by Salthouse, Berish, and Miles (2002) is used for this applet and explores these cognitive areas in more detail to evaluate the relationship between age and cognitive functioning.

Richard Alan Hullinger, Indiana University, Bloomington
Melanie Maggard, University of the Rockies

Statistical Lesson.

A topic related to correlation is validity. The text defines validity as the goodness with which a concrete event defines a property. It uses the example for concrete event as “frequency of smiling” to define the property of “happiness.” When we create a test to measure the property “statistical knowledge,” we want to make sure the test is valid in that it actually measures statistical concepts and skills, not attitudes, feelings, or other things. So, using the concept of correlation loosely, do the items in the test appear “on the face” to relate or correlate to what we are measuring? This is what we refer to as face validity.

There are two other types of validity that help us evaluate a test – convergent and discriminant validity.

  • Convergent validity

    tells us how well a test measures a specific property compared to another test that measures the same property. We expect the relationship between two tests that are measuring the same or a similar property to have a high correlation.

  • Discriminant validity

    tells us how well a test measures a property compared to another test that measures a different property. We expect the relationship between two tests that are measuring different things to have little or no correlation, meaning there is no relationship between what these tests measure.

Question 1

Salthouse, Berish, and Miles (2002) determined that certain tests of cognitive ability were correlated with each other and appeared to measure the same property. Those tests that measured similar properties had higher correlations or higher criterion validity with each other and those that measure dissimilar things had lower correlations or higher discriminant validity with each other. Knowing that we want the correlation to be as high as possible (closer to 1.00 than 0.00) for convergent validity, what test was most related or similar to block design? (Pick “Block design” for the X-axis. Cycle through the other tests on the Y-axis to see which one has the highest correlation (look at the r value) and has the tightest clustering of points on the graph to a straight line.)
Correct.
Incorrect. Spatial relations and paper folding had the highest correlations, both above 0.75, with block design.
1 / 10