• 0

# Interact: Variables, Measures, & SPSS Practice Worksheet

Related Documents
1 / 3100%
PSYC 510
VARIABLES, MEASURES, & SPSS PRACTICE WORKSHEET - SCALE OF
MEASUREMENT
Pracce 1 – Basic Concepts
Classify each operaonal variable below as nominal, ordinal, interval, rao (or scale).
a. Degree of pupil dilaon in a person’s eyes in a study of romanc couples (measured in millimeters)
b. Number of books a person owns.
c. A book’s sales rank on amazon.com
d. The language a person primarily speaks at home.
e. Naonality of the parcipants in a cross-cultural study of Canadian, Ghanaian, and French students.
f. A student’s le'er grade in school.
g. Narcissism as operaonally de(ned using the Narcissisc Personality Inventory (NPI – Raskin & Terry, 1988). This
40 queson survey counts the number of statements selected associated with narcissism and provides a single
score ranging from 0 – 40, with a higher score indicang greater levels of narcissism.
h. The queson “If you were given the chance, would you reapply to your current job?” Opons: Yes or No
i. The queson “Do you have fun at work?” using a sliding scale akin to Likert. People are shown all numbers 1 –
10 like in Likert, but instead of having to select a solid value, the parcipant can “slide” a bar anywhere on that
range.
Pracce 2 - Applicaon
Now that you can idenfy the scale of measurement, let’s pracce doing this from a scenario. For each scenario, idenfy
the independent and dependent variables, and the scale of measurement for each.
2-1.Does the level of seniority of a CEO a>ect a (rm’s performance? CEO level of seniority was categorized as New (0-1
years at the (rm); Moderate (2 – 7 years at the (rm), and Advanced (8+ years at the (rm). The (rm’s performance
was de(ned as the simple count of the number of acquisions in the last year.
2-2. This study examines whether there is a di>erence between CEO narcissism in male and female CEOs. Narcissism
was operaonally de(ned as the CEO’s use of (rst-person singular pronouns in interviews (determined by counng
the number of (rst-person singular pronouns divided by the sum of those pronouns plus all (rst- person plural
pronouns for a percent ranging from 0 – 100%).
Page 1 of 3
Review of Material: Scale of measurement – nominal, ordinal, interval/rao/scale (Module 2)
Informaon from: Video “Measurements” and Ch. 3 Jackson “Scales (Levels) of Measurement”
Addional pracce: Jackson ch. 3 Crical thinking check 3.1 Q2; MC self-test Q1 – 4 (answers in e-book for self-check)
Note: this is an OPTIONAL worksheet to pracce applying some of this module’s key concepts. Please make sure you
complete all assigned readings and watch this module’s presentaons before a'empng the worksheet! Try to
PSYC 510
ANSWER KEY for Scale of Measurement
Pracce 1 – Basic Concepts
Classify each operaonal variable below as nominal, ordinal, interval, rao (or scale). Answers are in red. Remember,
NOMINAL implies there are categories. ORDINAL requires magnitude. INTERVAL requires equidistant spacing between
levels of magnitude, and RATIO requires an absolute zero. These are ordered from least to most powerful. You go with
the scale of measurement that is the most powerful with all criteria met when labelling a variable’s scale of
measurement. Remember, for the purposes of this class (and primarily in the social sciences), Likert-based quesons will
always be considered interval (or SCALE). “SCALE” encompasses both interval and rao measurement types, as when it
comes to stascal tests, we don’t have to di>erenate.
a. Degree of pupil dilaon in a person’s eyes in a study of romanc couples (measured in millimeters) RATIO /
SCALE
b. Number of books a person owns. RATIO / SCALE
c. A book’s sales rank on amazon.com ORDINAL (has categories and magnitude, but it could be that the 3
rd
most
popular may have 50000 more votes than the 4
th
, whereas there is only 24 vote di>erence between the 2
nd
and
3
rd
… so not necessarily equidistant spacing).
d. The language a person primarily speaks at home. NOMINAL (no magnitude)
e. Naonality of the parcipants in a cross-cultural study of Canadian, Ghanaian, and French students. NOMINAL
(no magnitude)
f. A student’s LETTER grade in school. ORDINAL – much larger range for an F than any of the others, thus there is
no equidistant spacing. However, there is categories and magnitude.
g. Narcissism as operaonally de(ned using the Narcissisc Personality Inventory (NPI – Raskin & Terry, 1988). This
40 queson survey counts the number of statements selected associated with narcissism and provides a single
score ranging from 0 – 40, with a higher score indicang greater levels of narcissism. RATIO (since there is an
absolute zero) or SCALE
h. The queson “If you were given the chance, would you reapply to your current job?” Opons: Yes or No
NOMINAL (categories only)
i. The queson “Do you have fun at work?” using a sliding scale akin to Likert. People are shown all numbers 1 –
10 like in Likert, but instead of having to select a solid value, the parcipant can “slide” a bar anywhere on that
range. A sliding scale was described as akin to Likert – people are shown all numbers 1 – 10 like in Likert, but
instead of having to select a solid value, the parcipant can “slide” a bar anywhere on that range. They don’t
necessarily see the “inbetween” values, but the computer can provide the “exact” locaons. This gives more
range of data and so is typically considered more favorable / powerful by researchers in the debate of scale of
measurement. Regardless, in this course and typically in the social sciences, Likert is considered “scale” (or
“interval”), so that would be this scale of measurement as well. (As long as data is interval / rao / scale you can
run the same stascs, but it is good to know the theorecal di>erences! Remember – interval doesn’t have a
true zero; rao does. “Scale” is used to encompass both interval and rao data).
Pracce 2 - Applicaon
Now that you can idenfy the scale of measurement, let’s pracce doing this from a scenario. For each scenario, idenfy
the independent and dependent variables, and the scale of measurement for each.
Page 2 of 3
PSYC 510
2-3.Does the level of seniority of a CEO a>ect a (rm’s performance? CEO level of seniority was categorized as New (0-1
years at the (rm); Moderate (2 – 7 years at the (rm), and Advanced (8+ years at the (rm). The (rm’s performance
was de(ned as the simple count of the number of acquisions in the last year.
Remember “Independent” and “dependent” variables were covered last module. In this scenario, they want to see if
level of seniority a>ects a (rm’s performance. The independent variable (“cause”) then, is level of seniority. This is
actually a subject variable (wasn’t manipulated). Regardless, the scale of measurement is ORDINAL because there is
categories and magnitude, but no equidistant spacing within groups (new covers 24 month period; moderate is 5
years; advanced is an even larger range). Dependent variable is the “e>ect” and would be “(rm’s performance”.
Number of acquisions is RATIO (or scale), as there can be zero, and there is equidistant spacing.
2-4. This study examines whether there is a di>erence between CEO narcissism in male and female CEOs. Narcissism
was operaonally de(ned as the CEO’s use of (rst-person singular pronouns in interviews (determined by counng
the number of (rst-person singular pronouns divided by the sum of those pronouns plus all (rst- person plural
pronouns for a percent ranging from 0 – 100%).
Since again nothing is manipulated, to (nd out what the independent (subject) variable is, you ask what they think is
the “cause” vs. “e>ect”. In this scenario, they think gender may help explain narcissism. So, the independent
variable is gender, and the dependent variable is narcissism. Gender is NOMINAL (only categories). Narcissism is
RATIO / SCALE (percent range from 0 – 100% so there is a zero and equidistant units because each).
NOTE: There are sll opportunies for extra pracce! Try compleng Jackson ch. 3 Crical
thinking check 3.1 Q2 and MC self-test Q1 – 4 (answers in e-book for self-check)
Page 3 of 3