RE SEARCH ME THO DOLOG Y MODULE 3 – PART B MEAS URE MENT
MEASUREMENT Selecting
observable empirical events
Using
numbers or symbols to represent aspects of the events
Applying
a mapping rule to connect the observation to the symbol.
Wh at is M easured? Objects: • Things of ordinary experience • Some things not concrete Properties: • Characteristics of objects.
Characteristics of Data Classification Order Distance
(interval between numbers) Origin of number series.
Data Types Order
Interval
Origin
Nominal none
none
none
Ordinal
yes
unequal
none
Interval
yes
none
Ratio
yes
equal or unequal equal
zero
Nom inal Dat a / S cal e * **
Basic Characteristics
Examples
Descriptive
Numbers identify & classify objects
Store types, gender, members in teams
Percentages, Mode
Chi-square, Binomial test
Inferential
Example - Nominal
Runner 7
Runner 11
Runner 3
Ordina l D ata / S ca le * **
Basic Characteristics
Examples
Descriptive
Inferential
Numbers indicate relative position of objects but not magnitude of difference
Quality rankings, market position, social class
Percentages, Median
Rank-order Correlation, Friedman ANOVA
Example - Ordinal
Third Place
Second Place
First Place
Basic Characteristics
Differences between objects can be compared; zero point is arbitrary
Examples
Temperature, Attitudes, Opinions, Index numbers
Descriptive
Range, Mean, Standard Deviation
Inferential
Correlation, t-test, ANOVA, Regression, Factor Analysis
In te rva l D ata / Scale ** *
Example - Interval
Rating 8.2
Rating 9.1
Rating 9.6
Ratio D ata / S cale ** *
Basic Characteristics
Examples
Zero point is fixed; Ratios of scale values can be computed
Length, Weight, Age, Income, Costs, Sales
Geometric & Harmonic Mean
Coefficient of variation
Descriptive Inferential
Example - Ratio
15.2 seconds
14.1 seconds
13.4 seconds
Sources of Meas ur ement Diff erenc es Respondent Situational
factors Measurer or researcher Data collection instrument.
Characteristics of Sound Measurement Validity
– Extent to which a test measures what we actually wish to measure Reliability – Accuracy and Precision of a measurement procedure Practicality – Wide range of factors of economy, convenience and interpretability.
Validity in Experimentation *** Internal
Validity: Measure of accuracy in an experiment Measures whether the manipulation of the independent variables, or treatments, actually caused the effects on the dependent variable (s).
Validity in Experimentation ***
External
Validity: Determination of whether the cause-and-effect relationships found in the experiment can be generalised.
Validity Content
Validity
Criterion-Related
•Predictive •Concurrent
Construct
Validity
Validity.
Conte nt V alid ity Degree
to which the content of items adequately represent the universe of all relevant items under study Sometimes called face-validity *** Methods:
• Judgemental • Panel Evaluation.
Criterio n-R ela te d Va lid ity Degree
to which predictor is adequate in capturing the relevant aspects of the criterion
Method:
• Correlation.
Conc urre nt Va lid ity
Description of the present; Criterion data are available at same time as predictor scores.
Pred ic tiv e Va lid ity
Prediction of the future; criterion data are measured after the passage of time.
Cons tru ct Va lid ity
What accounts for the variance in the measure? Identifies the underlying constructs being measured and how well the test represents it Methods: Judgemental; Correlation; Factor Analysis; Multi-variate Analysis.
Relia bility Stability
(Test-retest)
Equivalence Internal
(Parallel forms)
Consistency (Split-half, KR20,
Cronbach’s alpha).
Tes t-R ete st (Sta bil ity ) Same
test is administered twice to same subjects over an interval of less than six months
Method:
Correlation.
Stability means one can secure consistent results with repeated measurements with same instrument.
Bias Caused By...
Time delay between measurements Insufficient time between measurements Respondent thinks there is disguised purpose Respondent forms new opinion before retest (Topic sensitivity) Introduction of moderating variables between measurements.
Parallel Forms (Equivalence) Test
administered simultaneously or with a delay
Method:
Correlation.
Equ iv ale nce
Concerned with variations at one point in time among observers and samples of items
Interrater Reliability
Delayed equivalent forms (to prevent “order of presentation” effect).
Cr onb ach’s Alph a (Int ernal Con sis ten cy )
Degree
to which instrument items are homogenous and reflect the same underlying constructs
Method:
Specialised Correlation formulae.
Improving Reliability Minimise External sources of variation Standardise conditions under which measurement occurs Improve investigator consistency Broaden measurement questions Add more observers or occasions Improve internal consistency.
Practicality
Economy
Convenience
Interpretability.