Dr.

Zak Case Study 1

Running Head: DR. ZAK CASE STUDY

Dr. Zak Case Study

Valentina Akyol

Courtney Scott

Nellie Rodriguez

Tiffany Christie

LaKeesha Carroll

University of Phoenix

Psychology 525: Measurements and Statistics

Dr. Matthew D. Geyer

April 4, 2009
 Dr. Zak Case Study 2

Dr. Zak Case Study

A case study was developed by a man named Dr. Zak. He developed a test to measure

depression among college students. Dr. Zak developed five questions for each student to answer

in order to measure depression. The scale of measurement Dr. Zak used was the nominal scale

of measurement. In this case study Team D will discuss if the nominal scale is the best

measurement Dr. Zak could have used and why. Team D will also discuss if he used good

samples in which to norm his test, and what other items could have been included. Team D will

begin with discussing standard deviations of Dr. Zak study.

Standard Deviations and Standard Scores

Dr. Zak used a total of 100 university students in his case study. They were all instructed

to take a test made up of 5 questions. His samples included 30 men, and 70 women. Out of this

group four were African American, six were Hispanic, and one person was Asian. Dr. Zak used

a mean of 3.5 on his test with a standard deviation of .5. According to Cohen and Swerdlik

(2005), “standard deviation can be defined as a measure of variability equal to the square root of

the average squared deviations about the mean. More succinctly, “it is equal to the square root of

the deviance” (p. 80). Several steps have to be taken in order to receive standard deviation

scores. The six steps consist of: compute the mean for the data set, compute the deviation by

subtracting the mean from each value, square each individual deviation, add up the squared

deviations, divide by one less than the sample size, and the final step is take the square root

(What is Standard Deviation, 2002, ¶ 1). Team D used these calculations in order to come up

with Sally’s scores.

 Dr. Zak Case Study 3

Sally scored a 1.5 on the test which means her standard deviation is a 4 which can be

computed mathematically as, (3.5-1.5)/.5 = 2/.5 = 4. Another student named Billy had a standard

score of 5. There were five questions on the test. Yes equals 1 and No equals 0. In order for him

to score 5, he had to have answered yes on all 5 in which he did. Here are Billy’s scores

computed mathematically. (5-3.5)/.5 = 1.5/.5 = 3 three standard deviations above mean and 3.5

+ 3(.5) = 3.5 + 1.5 = 5 mean plus three standard deviations is 5.

Dr. Zak Scale of Measurement Used

In order to measure these Dr. Zak had to use a certain scale of measurement called the

nominal scale of measurement. The nominal scale is the simplest form of measurement. The

nominal scale measures a particular characteristic or characteristics and places them in "mutually

exclusive or exhaustive categories" (Cohen and Swerdlik, 2005, p. 54). This form of

measurement is appropriate for Dr. Zak's case because Dr. Zak uses "yes" or "no" responses for

his questions. Yes or no answers are very common in nominal scaling. In each of Dr. Zak's

questions an answer of "yes" or "no" results in placement of a "mutually exclusive" category:

depressed or not depressed, sad or not sad, changes in eating and sleeping or no changes, life is

getting better or not, and happy most of the day or not. A "yes" answer is assigned a score of

"1," and a "no" answer is assigned a score of "0." The numbers are used solely for classification

purposes. They are not added or subtracted. So, a "yes" answer with a score of "1" does not

make it a better answer than a "no" with a score of "0" (Cohen and Swerdlik, 2005, p. 64-65).

Since Dr. Zak's case is a very simple one, any other form of scaling would not seem

appropriate. For example, "ordinal" scaling would not be appropriate simply because it allows

"classification." Dr. Zak's test does not require that participants be classified or "rank-
 Dr. Zak Case Study 4

ordered" on one of the characteristics. It simply puts the participant into a "mutually exclusive"

category. In this case, depressed or not depressed. Finally, it is not clear how Dr. Zak will

measure his test. As with any psychological tests, Dr. Zak's test should be reliable and valid.

Cassel (2003) stated that "the validity of any psychological test begins with the credibility of the

Test Taker (p.1). Perhaps Dr. Zak should base his findings on the DSM-IV (1994). "It consists

of a five point rating scale where a rating of 5 suggests full global functioning; a rating of 3 as

average; and a rating of 1 as the absence of global functioning. The Global Assessment

Functioning Scale was determined to be critical in the health and success of all individuals, and

is being used for such purposes" (Cassel, 2003, p.1).

Dr. Zak samples on which to norm his test

When developing a test, the developer defines a group of people who is representative of

the population for the design of the test. The population is "the complete universe or set of

individuals with at least one common, observable observation" (Cohen and Swerdlik, 2005, p.

103). Dr Zak's sample included 100 university students. Four of the students were black, six

were Hispanics, and one was Asian. Although, it seems like this group is a good one to sample,

more information is needed to obtain good results on his test. Therefore, Dr. Zak does not have a

good sample on which to norm is test. For Dr Zak to have a good sample on which to norm his

test, he would have to consider other things such as age norms, grade norms, and national

norms. Age groups include different samples of test-takers, at various ages at the time each

student took the test (Cohen and Swerdlik, 2005, p.107). Dr. Zak does not provide the students

ages in his sample. Grade norms indicate at which level of education each student is at the time

he or she took the test (Cohen and Swerdlik, 2005, p. 107). Dr. Zak's sample consists of
 Dr. Zak Case Study 5

university students. Dr. Zak did not state if they are freshman, sophomores, juniors, or seniors.

National norms "are derived from a normative sample that was nationally representative of the

population at the time of the norming study was conducted" (Cohen and Swerdlik, 2005, p.

108). In psychology, this would be obtained by testing a large group. This would include

"different variables of interest such as age, gender, racial/ethnic background, socioeconomic

strata, geographical location, and different types of communities within the various parts of the

country" (Cohen and Swerdlik, 2005, p. 108). The information provided to us did not include

the student's age, socioeconomic strata, geographical location, or the type of community each

student came from. This information would have been helpful and would have made Dr.

Zak's sample a good sample on which to norm his test.

Dr. Zak’s Domain Sampling

Cohen and Swerdlik (2005) states"in the process of developing a test, a test developer has

targeted some defined group as the population for which the test is designed" (p. 103). In Dr.

Zak's test his domain sampling consisted of the 100 university students who were African

American, Hispanic, Asian, and others. He also sampled male and female but most were females

he targeted. Dr. Zak could have expanded his sampling by using people who were poor, rich,

had jobs or no jobs, and those of different religions. Dr. Zak could have also used different

subgroups for his testing. According to Cohen and Swerdlik (2005), "subgroups within a defined

population may differ with respect to some characteristics, and it is sometimes essential to have

these differences proportionately represented in the sample" (p. 103). Dr. Zak could have

targeted people who live in the New York area which is known as stratified sampling (Cohen and
 Dr. Zak Case Study 6

Swerdlik, 2005, p.103). These are all different examples of different domains Dr. Zak could have

included in his sampling.

Changes which could have Improved Dr. Zak Test

Dr. Zak’s study should use a broader ranger of participants. The number of men

and women should be better proportioned as well as the cultures and ages. Surveying people the

same age would present better data to record. It is important to have more equal distributions of

culture because people from different cultures may have different assumptions about the

questions being asked or interpret them in different ways (Cohen, R.J. & Swerdlik, M.E., 2005).

There should be more questions added to the survey. This would provide a better detail of the

person for the disorder. Dr. Zak is on the right path by comparing his inventory with the BDI,

because it is important to be knowledgeable about other tests to fulfill his objectives (Cohen, R.J.

& Swerdlik, M.E., 2005). However, because his results did compare to the results to the BDI

test, Dr. Zak's test is not reliable or has validity. DR. Zak needs to makes sure his test is reliable.

His measurements need to be consistent: "In addition to being reliable, tests must be reasonably

accurate. In the language of psychometrics, tests must be valid" (Cohen, R.J. & Swerdlik, M.E.,

2005). Another way Dr. Zak could have improved his test is by preparing a rough draft of the test

which could provide an idea of what test items need to be included in the final version, which

will discriminate between test takers. Cohen and Swerdlik (2005) state "more test items are

needed to adequately sample the domain of the test" (p. 201). Overall the test works well but it

needs some improvements.

 Dr. Zak Case Study 7

Dr. Zak and the Beck Depression Inventory

A correlation of r=.14 indicates a positive relationship of small magnitude between the

two tests. Being positive, the value of r specifies a direct correlation in which the two variables,

the tests, simultaneously proceeded in the same direction, and with such a low value, it could be

inferred that both the variables decreased. Furthermore, values of +1 or -1 signify a perfect

correlation between variables and +0 or -0 signify no relationship at all, and since Dr. Zak's test

and the BDI have a correlation value of .14, there is almost no correspondence between the

tests (Cohen and Swerdlik, 2005).

The low correlation value between the two tests can be suggestive of several things. The

first is the degree of connected variation between the variables. Having a low correlation results

in high levels of variation, which in turn decreases the predictability of one variable's value when

the other is known (Cohen and Swerdlik, 2005). Secondly, the two tests have different

approaches to the measurement of depression. The two instruments are not a dependently

operating system, and instead, are two unrelated measurements of depression. The last inference

that can be made from the correlation value is that of the possibility of error. Whether it was the

administration of the tests, the analysis, or the test takers themselves, something occurred

that caused the two instruments to have little magnitude of co-relation (Cohen and Swerdlik,

2005).

In conclusion, Team D has determined several things from Dr. Zak case study on

depression. Team D determined Sally’s deviation scores and Billy’s standard scores. Dr. Zak

used the nominal scale which was appropriate because his questions were either yes or no

questions. These yes or no questions are common in nominal scaling which Team D concluded
 Dr. Zak Case Study 8

was the appropriate scale for Dr. Zak to use. Team D also concluded Dr. Zak needed more

information in which to obtain better results on his test. It was concluded he should have

considered other things such as age norms, grade norms, and national norms. The team

determined the correlation between Dr. Zak test and the BDI had almost no correspondence

between the two tests. It was determined Dr. Zak had several areas in his test that were good but

other areas needed improvement.

 Dr. Zak Case Study 9

Reference

Cassel, R. (2003, September). Confluence is a primary measure of test validity and it includes

the credibility of test taker. College Student Journal, 37(3), 348. Retrieved April 4, 2009,

from Academic Search complete database.

Children's Mercy (2002). What is a standard deviation?. Retrieved April 4, 2009, from

http://www.childrensmercy.org/stats/definitions/stdev.htm

Cohen, R. J. and Swerdlik, M.E. (2005). Psychological testing and assessment: An introduction

to tests and measurement (6th ed.). New York: McGraw Hill.