Sie sind auf Seite 1von 52

Measurement Scales

Definition

• Measurement is to discover the extent,


dimension, quantity or capacity of
something, especially by comparison
with a standard.

• Measurement in research consists of


assigning numbers to empirical events
in compliance with the set of rules.
Three parts of measurement

• Selecting a observable empirical events


• Developing a set of mapping rules: a
scheme for assigning numbers or
symbols to represent aspects of the
event being measured.
• Applying the mapping rule(s) to each
observation of that event.
Goal of measurement

• To provide the highest quality, lowest


error data for testing hypotheses.
Must know…..

• Concept
• Construct
• Constitutive Concept
• Operational Concept
Characteristics of a Good
measurement

• Unidimensionality
• Linearity
• Validity
• Reliability
• Accuracy and precision
• Simplicity
• Practicability
Investigative
Que
Moving from
What do I plan to do with the data
Investigative
Description? Que to
Exploration? Measurement
Discovering of difference?
Que.
Finding of relationships?

Is dn Type of data? How many


expected to groups will be
be normal? Nominal compared?

What is my Ordinal Are groups


expected related or
sample size? Interval independent?
Ratio

What Measurement Scale should I use?

Measurement
Que
Measurement Scales

• Nominal scale
• Ordinal Scale
• Interval Scale
• Ratio Scale
Nominal Scale

• Most elementary method of


measurement.

• Categorical data and numbers that are


simply used as identifiers or names.

• Mutually exclusive and collectively


exhaustive.
Nominal Scale Cont…

• Numbers doesn't have any quantitative


value.

• Arithmetic opn. – Counting

• Mode is the only measure of central


tendency.
Nominal Scale Cont…

Examples:
Numbers on Cricket players' jerseys
that are used to identify each player.

Classifying data, e.g. m/f

No ordering, e.g. it makes no sense to


state that M > F

Arbitrary labels, e.g., m/f, 0/1, etc


Ordinal Scale

• Numbers, letters or numerals are


used to rank objects.

• Relative position of two or more


objects on some characteristic.

• Used in Customer-oriented research.


Ordinal Scale Cont…

• Doesn’t provide how much less or more the


attribute is.

• size of the interval between any two


numerals is unknown.

• Operations: Mode, Median, Non parametric


tests.
Ordinal Scale Cont…

Examples:
Grades scored by students at IBS.
Rank preferences for several brands,
flavors.
Restaurants ranked from most like to
least liked.
Movie Ratings.
Interval Scale

• Also uses numerals to rank objects.

• Numerically equal distances on the


scale represent equal distances in the
property being measured.

• Arbitrary zero i.e. no natural zero.


Interval Scale Cont…

• Descriptive measures: mean, median,


mode, range and standard deviation.

• Bi-variate correlation analysis, t-test,


analysis of variance tests.
Interval Scale Cont…

Examples:
Centigrade and Fahrenheit temperature scales.
Measurement of Sea Level.
Personality measures.

Differences make sense, but ratios do not


(e.g., 30°-20°=20°-10°, but 20°/10° is not
twice as hot!)
Ratio Scale

• All properties of Interval.


• True, Natural or Absolute Zero.
• Value of zero on a ratio scale
indicates the complete absence of
the characteristic of interest.
• All Descriptive measures and
inferential techniques are applicable
to ratio-measured data.
Ratio Scale Cont…

Examples:
Length, time, force, volume and area.
A negative length is not possible.
Temperature in Kelvin.
Quantitative data such as interval or
ratio data can be converted to
categories on a qualitative scale but not
Vice Versa.
Fundamental Difference

Indicates Indicates direction Indicate Absolute


Difference of Difference amount of Zero
Difference
Nominal 

Ordinal  

Interval   

Ratio    
Permissible Arithmetic Operations

Nominal Ordinal Interval Ratio

Counting Greater Addition or Multiplication


than or Subtraction or Division of
Smaller of scale scale values
than values
Examples of Appropriate Statistics

Nominal Ordinal Interval Ratio

Chi Square Median, Mean, SD, Coefficient


Runs test Interquartile t –test, of Variation
range, ANOVA,
Spearman's rank Regression
correlation,
Kolmogorov-
smirnov test,
Mann-Whitney U
Reliability

• Meaning:
• The extent to which the instrument
yields the same results on repeated
trials.
• Consistency
• Free from error
Reliability Cont…

• Examples:
• ordinal measures are reliable if they
consistently rank order items in the same
manner
• reliable interval measures consistently rank
order and maintain the same distance
between items.
Concept of Reliability

• Two Dimensions
• Repeatability
• Internal Consistency
Tests of Reliability

• Retest Method

• Same test is given to the same people after


a period of time.
• Examination of consistency
• Eg. Measuring Satisfaction level
Tests of Reliability Cont…

• Problems:
• Time delays between measurement
• Insufficient time between measurements
• Respondent’s discernment of a disguised
purpose
• Topic sensitivity
• Introduction of extraneous moderating variables
between measurements
Equivalent Form Method

• Two measurement scales of a similar


nature are developed

• Measure the correlation of scores


generated by two instruments.
Equivalent form Method Contd..

• Problem
• Framing two totally equivalent
questionnaires
Internal Consistency method

• Testing whether the data give same


results even after some manipulations.

• Example
Internal Consistency method
Contd..

• Problem:
• Dependent on data manipulation or
divisions
Cronbach’s Alpha

Where
N = Number of items
= Sum of item variance
= Variance of the total composite
Kuder Richardson Formula-20

• In case of difficulty in obtaining the


data at equal interval of time
Validity

• Ability of a scale or a measuring instrument


to measure what it is intended to measure

• Does the instrument really measure what


its designer claims it does?
Example

• Validity of an exam

• Measure of the morale of employees

based on their absenteeism alone


Classification of Validity

• Face Validity

• Collective agreement of experts and


researchers on the validity of the
measurement scale
• Experts determine if the scale is measuring
what it is expected to measure
Classification of validity

• Content Validity
• Adequacy in the selection of relevant
variables for measurement
• Measurement contains a
representative sample of the universe
of subject matter of interest : Content
Validity is good
Classification of validity

• Determination of content validity


• Intuitive logical process by research
designer
• Use a panel of persons to judge how
well the instrument meets the
standards
Classification of validity

“ Content Validity is primarily


concerned with inferences about
test construction rather than the
inferences about the test scores ”
Classification of validity

• Criterion Related Validity


• Degree to which a measurement
instrument can analyze a variable
• If a new measure is developed, it must
correlate with other measures of the
same construct
Classification of validity

• Criterion related validity can be


categorized as :
• Predictive Validity
• Concurrent Validity
Classification of validity

• Predictive validity
• The extent to which a future level of a
criterion variable can be predicted by a
current measurement on a scale
Classification of validity

• Concurrent validity
• Related with the relationship
between predictor variable and
criterion variable ; evaluated at the
same point of time
Classification of validity

• Construct Validity
• Degree to which a measurement
instrument represents and logically
connects through the underlying
theory
Classification of validity

• Example
• A particular product purchased by a
consumer or not, is not the question ;
why he has or has not purchased the
product is taken into account to judge
construct validity
Classification of validity

• Construct Validity is classified as


• Convergent validity
• Discriminant validity
Classification of validity

• Convergent Validity
• Extent of correlation among different measures that
are intended to measure the same concept

• Discriminant Validity
• Denotes the lack of or low correlation among the
constructs that are supposed to be different
Classification of validity

• Sensitivity
• Ability to accurately measure variability in
responses
• Generalization
• Amount of flexibility in interpreting the data in
different research designs.
• Relevance
• Appropriateness of using a particular scale for measuring variable.
Understanding Reliability and Validity
Sources of measurement problems

RESPONDENT ASSOCIATED ERRORS:


Non-Response Errors
It includes,
• Failure to respond completely
• Failure to respond to one or more questions of the surveyor
• Unit non-response, Item non-response
Response bias
Sources of measurement problems

• INSTRUMENT ASSOCIATED ERROR

• Poor questionnaire design

• Improper selection of samples

• Complex questionnaire
Sources of measurement problems

• SITUATIONAL ERRORS
• Example: Location of the interview

• MEASURER AS ERROR SOURCE


• Body language and gestures
• Failure to record the full response
• Using irrelevant statistical tools

Das könnte Ihnen auch gefallen