A Jeanne Site

## Statistics Site

California State University, Dominguez Hills
University of Wisconsin, Parkside
Latest update: February 4, 1999
Faculty on the Site.

## Lecture Notes on Exercise 2: Measurement

Jeanne at jcurran@csudh.edu
Subject line: stex02: measurement
First message line: Your name and class.
Second message line: Name of each member of your collaborative group. Body of message: xxxxxxx

Source materials for the following questions will be found in Adventures in Criminal Justice ResearchDowdall, Babbie, and Halley, Chapter 3, pp. 19-26.

Try to answer in 25 words or so. Make each answer integral, so that I can read it without reference to the exercise or the question itself.

1. What is validity? (DBH, at p.19)

Validity comes from "valid," meaning, effective, correctness or soundness of reasoning, correct, making good sense. It's a little different from "true," because true implies the opposite of false, leading us to think in dichotomies. Many answers can be valid, effectively reasoned and well-grounded.

In statistics, validity means that the measure "really measures" what it claims to measure. In most social science issues that is arguable. For example, IQ has face validity, validity based on your common sense. Asking lots of questions that would seem to measure how much you know about subjects we study in school would seem to measure your intelligence. But some people do very well on tests and not so well in real life. So how valid are our tests as measures of intelligence?

2. Discuss the validity of grades. (DBH, at p. 20, "workable indicators")

Validity of grades: Grades are generally taken to be a measure of intelligence. The higher a student's grades, the smarter the student is presumed to be. But there are validity problems. Some people do well on tests, but that does not translate into competent performance on real tasks.

Validity will vary depending on the strength of the connection between the subject being graded and the practical skills that test purports to measure. For example, elementary school grades would not be valid as measure of legal performance. The connection is too general, too distant. But law school grades might be a more valid measure of legal performance. Might be. All of this is probabilistic. If a student gets good grades he/she will probably perform better on tasks related to those grades.

3. The statement that age is a ratio variable is arguable. Why? (statement on p. 23 of DBH. Answer in lecture.)

Age satisfies the arithmetic conditions for a ratio variable, since there is zero age, and since each year is the same on the calendar. But what about the growth in potential between 1 and 4? Is that matched by three years between 27 and 30? What about 48 to 51? What about 87 to 90? Arithmetically these three year periods may be the same. But are they the same psychologically, sociologically, and biologically? What about years associated with ritual promotions? Like 18 or 21? What about 65? What about 30? 40? 50? and 60? What about puberty? retirement? All these questions force us to question the ratio nature of age in any but arithmetic terms

4. What are mutually exclusive categories? (DBH, at p. 24)

Mutually exclusive categories are categories which do not overlap, so a respondent who fits into one cannot fit into any other.

Samples:

• gender: male, female
• education: elementary, high school, post high school
• party id: democrat, republican, other

5. What is an informed respondent? (DBH, at p. 25)

An informed respondent is someone whose professional work or whose membership in a given group puts him/her in a position to be able to answer questions for the group he/she belongs to or works with. DBH give the example of administrators at a college answering surveys for the whole college. Parole workers may answer questions about their clients. They have special information that gives them this broad knowledge.