FRHD 3070 NEW FINAL EXAM WITH ALL QUESTIONS AND CORRECT
ANSWERS A+ GRADED
Pros of using existing measures - Answer save time, money, can make comparisons with
other results
Cons of using existing measures - Answer harder to cover abstract content, not
appropriate for target pop., format not appropriate for collection mode and different
researchers have different views
operational definitions - Answer sequence of steps or procedures that a researcher
followed to obtain a measurement (concrete and measurable)
conceptual defintion - Answer very basic definition of what something is - in order for it
to work it research it must be operationalized
components of observed score - Answer observed score = function of true score +
systematic error + random error
Measure twice and cut once (analogy) - Answer ways to minimize measurement error (it
is always present)
a variable has three components - Answer a true score, systematic error and random
error
true score - Answer measures construct of interest
systematic error - Answer things that are the same over and over again but also
measure constructs of disinterest
,random error - Answer constantly ever changing, nothing methodical about it
low and high levels of measurement - Answer lower is less sophisticated and higher is
more sophisticated
4 levels of measurement - Answer nominal, ordinal, interval, ratio
nominal - Answer measuring at categorical measure, they are mutually exclusive (sex,
religion, ethnicity) - no hierarchy
ordinal - Answer identifies differences among categories -and ranks categories ranging
from low to high (ranked) on a continuum, (report card grades, likert scales)
interval - Answer identifies differences among categories and ranks categories and
intervals between adjacent categories and each interval is spaced out equally, no true
zero (temperature)
ratio - Answer identifies differences among categories and ranks categories, measures
distance between categories and has a true zero, meaningful, total absence of
characteristic (income, years of education, cigarettes - if you don't smoke any you
smoke zero) (money or age)
nominal stats - Answer frequencies (how many people endorse a certain category)
ordinal stats - Answer frequencies; *median* (because ranking)
interval & ratio stats - Answer frequencies; median and *mean* (the average)
what determines the types of mathematics you can do with data - Answer levels of
measurement
,levels of measurement analogy - Answer like a nesting doll because you move from
lowest to highest level (nominal to ratio)
example of nominal in marathon - Answer numbers on runners
example of ordinal in marathon - Answer place in race
examples of interval /ratio in marathon - Answer amount of time taken to complete race
Correlation - Answer stat measure of strength association between 2 variables
+1.00 correlation - Answer high scores of one variable are associated with high scores
of other variable and vice versa
-1.00 correlation - Answer high scores on one variable are associated with low scores of
other variable (low self-esteem associated with drug use)
zero correlation (points scattered) - Answer SES and honesty - no relationship between
the variables
effect size - Answer r (correlation) summarizes strength of relationship
r = 0.10 - Answer effect size is small (weak) and range is less than 0.30
r = 0.30 - Answer effect size is moderate (medium) and range is 0.30 - 0.49
r = 0.50 - Answer the effect size is large (strong) and range is 0.50 & above
Reliability - Answer degree to which observed scores are free from errors of
, measurement
types of reliability - Answer test retest; alternative-forms; split-half; internal consistency;
and inter-rater
test-retest reliability - Answer the same measure conducted at 2 different times
rule of thumb for test re-test reliability - Answer the reliability coefficient of 0.70 or more
is an adequate
challenges to test-retest - Answer time passed between 2 administrations; interpreting a
low correlation; interpreting a high correlation; measuring people twice
alternative forms reliability - Answer 2 different forms during ONE occasion
split-half reliability - Answer single administration; split items and compare them - they
should be correlated because measuring the same construct
challenges to split half - Answer many different ways to split an instrument (odd VS
even); have different coefficient or vales depending on split
internal consistency reliability - Answer item homogeneity; Cronbachs coefficient alpha;
one administration, correlation between ALL items; 0.70 or higher suggests internal
consistency
Cronbach's coefficient alpha - Answer only used if construct is unidimensional - if multi
then should calculate it for each scale (intrinsic VS extrinsic)
inter-rater reliability - Answer consistency among different observers using the same
instrument
ANSWERS A+ GRADED
Pros of using existing measures - Answer save time, money, can make comparisons with
other results
Cons of using existing measures - Answer harder to cover abstract content, not
appropriate for target pop., format not appropriate for collection mode and different
researchers have different views
operational definitions - Answer sequence of steps or procedures that a researcher
followed to obtain a measurement (concrete and measurable)
conceptual defintion - Answer very basic definition of what something is - in order for it
to work it research it must be operationalized
components of observed score - Answer observed score = function of true score +
systematic error + random error
Measure twice and cut once (analogy) - Answer ways to minimize measurement error (it
is always present)
a variable has three components - Answer a true score, systematic error and random
error
true score - Answer measures construct of interest
systematic error - Answer things that are the same over and over again but also
measure constructs of disinterest
,random error - Answer constantly ever changing, nothing methodical about it
low and high levels of measurement - Answer lower is less sophisticated and higher is
more sophisticated
4 levels of measurement - Answer nominal, ordinal, interval, ratio
nominal - Answer measuring at categorical measure, they are mutually exclusive (sex,
religion, ethnicity) - no hierarchy
ordinal - Answer identifies differences among categories -and ranks categories ranging
from low to high (ranked) on a continuum, (report card grades, likert scales)
interval - Answer identifies differences among categories and ranks categories and
intervals between adjacent categories and each interval is spaced out equally, no true
zero (temperature)
ratio - Answer identifies differences among categories and ranks categories, measures
distance between categories and has a true zero, meaningful, total absence of
characteristic (income, years of education, cigarettes - if you don't smoke any you
smoke zero) (money or age)
nominal stats - Answer frequencies (how many people endorse a certain category)
ordinal stats - Answer frequencies; *median* (because ranking)
interval & ratio stats - Answer frequencies; median and *mean* (the average)
what determines the types of mathematics you can do with data - Answer levels of
measurement
,levels of measurement analogy - Answer like a nesting doll because you move from
lowest to highest level (nominal to ratio)
example of nominal in marathon - Answer numbers on runners
example of ordinal in marathon - Answer place in race
examples of interval /ratio in marathon - Answer amount of time taken to complete race
Correlation - Answer stat measure of strength association between 2 variables
+1.00 correlation - Answer high scores of one variable are associated with high scores
of other variable and vice versa
-1.00 correlation - Answer high scores on one variable are associated with low scores of
other variable (low self-esteem associated with drug use)
zero correlation (points scattered) - Answer SES and honesty - no relationship between
the variables
effect size - Answer r (correlation) summarizes strength of relationship
r = 0.10 - Answer effect size is small (weak) and range is less than 0.30
r = 0.30 - Answer effect size is moderate (medium) and range is 0.30 - 0.49
r = 0.50 - Answer the effect size is large (strong) and range is 0.50 & above
Reliability - Answer degree to which observed scores are free from errors of
, measurement
types of reliability - Answer test retest; alternative-forms; split-half; internal consistency;
and inter-rater
test-retest reliability - Answer the same measure conducted at 2 different times
rule of thumb for test re-test reliability - Answer the reliability coefficient of 0.70 or more
is an adequate
challenges to test-retest - Answer time passed between 2 administrations; interpreting a
low correlation; interpreting a high correlation; measuring people twice
alternative forms reliability - Answer 2 different forms during ONE occasion
split-half reliability - Answer single administration; split items and compare them - they
should be correlated because measuring the same construct
challenges to split half - Answer many different ways to split an instrument (odd VS
even); have different coefficient or vales depending on split
internal consistency reliability - Answer item homogeneity; Cronbachs coefficient alpha;
one administration, correlation between ALL items; 0.70 or higher suggests internal
consistency
Cronbach's coefficient alpha - Answer only used if construct is unidimensional - if multi
then should calculate it for each scale (intrinsic VS extrinsic)
inter-rater reliability - Answer consistency among different observers using the same
instrument
