PSYC 201 – Lab assignment
Correlation matrices, reliability, and validity
1. The table below contains the correlations among the six items that make up a questionnaire. Each question asked subjects to indicate the degree to which they agreed or disagreed with a statement. A 7-point Likert scale was used for each item. The higher the score, the more they agreed with the statement. Fifty subjects filled out the questionnaire.
Attitudes towards graduate school scale
a. I intend to apply to graduate school in my field.
b. Graduate school would make me more competitive to get the kind of job I want.
c. I would get a lot out of attending graduate school
d. I have no intention of attending graduate school.
e. I can get the kind of job I want without attending graduate school.
f. I am very interested in psychology, in general.
a b c d e f
a 1.0 +.6 +.5 -.8 -.7 +.15
b 1.0 +.7 -.6 -.7 +.2
c 1.0 -.5 -.5 +.1
d 1.0 +.8 -.2
e 1.0 +.08
f 1.0
a. Which of the correlations presented in the table above are significantly different from zero? Put a star beside the significant correlations.
b. In your opinion, is the questionnaire reliable in terms of its internal consistency? What does it mean for a questionnaire to be internally consistent? How is this different test-retest reliability?
c. Based on the data presented above, which, if any, items should be eliminated from the questionnaire? Why or why not eliminate one or more items?
2. The investigator has obtained total scores on the questionnaire used in Part 1. From each of 50 people they obtain the correlations among three questionnaires: the Attitude Towards Graduate School Scale (ATGSS) presented in Part 1 above, the Intention to Attend Graduate School Questionnaire (IAGSQ) , and the Scale of Attitudes Towards Golf (SATG). The investigator wants to use the correlations among these three questionnaires to evaluate the ATGSS in terms of its construct validity.
ATGSS IAGSQ SATG
ATGSS 1.0 +.85 +.12
IAGSQ 1.0 -.09
SATG 1.0
a. What does the term “construct validity” refer to?
b. Does the Attitude Towards Graduate School Scale have convergent validity? Why or why not?
c. Does the Attitude Towards Graduate School Scale have divergent validity? Why or why not?
3. A researcher has developed a new Brief Questionnaire on Depression (BQD) that consists of five statements rated on a seven-point Likert scale. The higher the rating, the more the participant agreed with the statement. Sixty-two people filled out the questionnaire. The items for the questionnaire are listed below:
a. I feel depressed.
b. I am happy most of the time.
c. I have had trouble sleeping lately.
d. I feel sad most of the time
e. I am very pleased with how things have been going lately.
Correlations among the five items are presented below:
a b c d e
a 1.0 +.04 -.2 +.12 -.07
b 1.0 -.7 -.9 +.9
c 1.0 +.6 -.7
d 1.0 -.8
e 1.0
a. If the investigator wants higher scores for the questionnaire to indicate higher levels of depression, which, if any, items will need to be reversed?
b. Which of the correlations presented in the table above are significantly different from zero? Put a star beside the significant correlations.
c. In your opinion, is the questionnaire reliable in terms of its internal consistency?
d. Based on the data presented above, which, if any, items should be eliminated from the questionnaire? Why or why not eliminate one or more items?
e. The investigator gives the BQD again to the same participants two weeks after the first administration. The correlation between scores for the BQD at Time 1 and Time 2 is +.2. Can the investigator claim that the BQD has test-retest reliability?
4. The investigator has obtained total scores on the questionnaire used in Part 1. From each of 62 people they obtain the correlations among three questionnaires: the Brief questionnaire on Depression (BQD) presented in Part 3 above, the Beck Depression Inventory (BD!), and the Baltimore Introversion/Extroversion Scale (BIES) (higher scores indicate that the person is more extroverted, lower scores indicate that the person is more introverted) . The investigator wants to use the correlations among these three questionnaires to evaluate the BQD in terms of its construct validity.
BQD BDI BIES
BQD 1.0 +.05 +.85
BDI 1.0 -.04
BIES 1.0
a. Does the Brief Questionnaire on Depression have convergent validity? Why or why not?
b. Does the Brief Questionnaire on Depression have divergent validity? Why or why not?
c. According to the correlation matrix among these three variables, what does it look like the Brief Questionnaire on Depression actually measures?