Sunteți pe pagina 1din 26

CHAPTER 16

Index, Scale and Procedure


for Special Measurements
INTRODUCTION
Combination of scores for a
measurement indicator will result a
new score known as index or scale.
Index-combination of two or more types
of measurement indicator scores.
Scale-more complex combination with a
specific pattern or arrangement.

(pg. 307 paragraph 1)


A. MEASUREMENT INDEX
Consists of a combination of several indicators.
May also consist of a combination of several sub-
indexes where each sub-index is formed from a
combination of several indicators.
It is better to form a measurement index using
sub indexes or indicators which are represented
by items.
Combining the aspects into an index will increase
the reliability ad validity of the research.

(pg. 307&308, Figure 16.2 & 16.3 )


B. ANALYSIS OF ITEMS TO FORM
A MEASUREMENT INDEX
Researcher must ensure that the item
representing each indicator are used in the
constructed measuring instrument to have
high reliability and validity.
These items must be analyzed and tested
before they are used on the research subjects.
Steps that should be taken:
i) give an accurate conceptual definition
refer to hypothesis concepts and clarify the using
operational concepts (measurement guideline)

(page 309, Figure 16.4)


ii. Form a table for determining
the sub-indexes
to ensure the items formed contain those
three sub-indexes.
number of items based on each sub is
listed.
each sub index may not necessarily have
same numbers of items as some sub-
indexes require more measurement items
to represent their concept.
each items must have high face validity-
represent the concept to be measured.
( page 310, table 16.1)
iii.Conduct a Pre-test
give the items to a group of subject from
the population of respondent that the
researcher want to study who were not
selected as respondents for the research
sample.
encourage the subjects to ask questions
if there are any doubts or uncertainties
about the measurement items.
improve problematic item and retype
them.
(pg. 310)
iv. Conduct a pilot test
- pilot test must be conducted on the
items by a group of subjects.
- data is analyzed to see if it is
reliable.
- method : test-retest, split-half(Pg.
method 310)
C. SPECIAL MEASUREMENT
SCALES
Measurement scale consists of
different combinations of scales of
measurements.
Standard scales used in social
sciences and education:
Likert scale
Thurstone scale
Guttman scale
Semantic differential scale (Pge.
311)
D. LIKERT SCALE
Rensis Likert, 1932
Basically uses 5 scales.
Advantages:
Easily managed and utilized.
Items can be easily answered by respondents.
Data collection has higher reliability.
Steps is forming Likert Scale
1. Developing proposed items
60-100 items
Based on theories/own understanding
(Pge. 312)
2. Evaluating the Suitability of the
Proposed items.
Given to panel of experts to evaluate the items,
yet their duty is not to answer the items. (refer
Figure 16.6)
Criteria:
Statement of items must be able to represent the
concept to be measured, consistent with operational
definition.
Avoid and
Form and structure must be standardized.
Avoid using two/more negative word in a statement.

3. Analysing the suitability of item.


Calculating the Inter correlation value
Calculating the Discrimination index
4. Placing items in Likert Scale
Errors in Likert Scale
Must be balanced at both end to
increase the validity and reliability of the
measurement. (Pge. 320, Figure 16.11)

Analysis of Likert Scale Data


Eg: differences between respondent
groups can be conducted using non-
paramedic test : Mann-Whitney U test
and Kruskal-Wallis H. Correlation
analysis: Spearman rho. Descriptive
analysis: frequency and percentage.
E. THE THURSTONE
SCALE
- Was built by Louis Thurstone who is
psychologist.

- This scale was built on his belief that human


beings can be assessed and evaluated on a
psychological basis.

- Contains statement representing an aspect of


the concept being measured, which are agree
and disagree

- The Thurstone scale is a uni-dimensional scale


used to measure a research concept.
1)Steps to forming the Thurstone
scale.
a)Developing proposed item statements.
-) Can form the item statements based on his understanding.

-) Ask other individuals for help with ideas on general


variables.

-) Discussion/ brainstorming session to generate ideas can be


conducted to produce proposed items (60 to 100 items)

-) These item statements will later be checked for suitability


on the following rules. ( refer page 321)
b) Evaluating the agreement level of
suggestion items.
- The suitability of each item must be evaluated before it
can be used.

- About 100 experts are needed to evaluate the items


based on a scale of 1 to11.

- The number 1 indicates most positive regarding the


concept and number 11 indicates most negative
regarding the concept
c) Analysis the suitability of proposed items.
- To analyse the suitability of each statement based on the
median value and the interquartile value for the selection
scores of the experts

- Refer page 323-327

d) Placing items in the Thurstone scale.


- The selected items will be randomly placed in the Thurstone
scale.

- Each item is followed by the choices agree and disagree.


2) Analysis of Thurstone scale data.
- Due to the Thurstone scale being an interval
scale, an analysis of differences between
respondent groups can be conducted using t-
test or ANOVA test while correlation analysis
can be conducted using Pearson r, and
descriptive analysis can be conducted using
the mean score and standard deviation
F. THE GUTTMAN SCALE

- Was built in 1940 based on Louis


Guttmans statement that when a
concepts items with different weightages
are arranged according to order, and if a
respondent answers positively towards
items with lighter weightages, he will
answer negatively for items with heavier
weightages based on his feeling.

- It is also known as the scalogram analysis


scale.
1)Steps to forming the Guttman
scale.
-) Is a uni-dimensional scale used to measure a
research concept.

-) It is developed based on its own procedure.

a)Developing proposed items.


-) Item statements can be formed by the researcher based on
his understanding of the concept (or by referring to theories
and models).

-) He can also seek assistance from others to generate ideas.

-) Produced statements must be able to represent the


different levels of weightage for the concept being studied.
b) Arranging statements according to
weightage.
- The weightage of the proposed statements will be
evaluated according to the concept tested.

- A group of experts (60 to 100) who are knowledgeable


about the concept will evaluate the weightage of these
statements.

c) Analysing suitability of items.


- Refer pages 330- 332

d) Placing items in the Guttman scale.

- The items selected will be placed in the Guttman scale


according weightage.
2) Analysis of Guttman scale
data.
- since the Guttman scale is an ordinal
scale, an analysis of the differences
between respondent groups can be
conducted using the Mann- Whitney
( difference between two group) or the
Kruskal- Wallis H test ( difference among
more than two groups).

- Correlation analysis can be conducted


using the Spearman rho correlation test,
whereas descriptive analysis can be
G. THE SEMANTIC
DIFFERENTIAL SCALE.
- Was built by Charles Osgood in 1957.

- It is a tri-dimensional scale and its strength is


that it allows a researcher to view the
respondents attitudes from various
perspectives.

- The box below shows two respondents who


gave connotations for the word bear ( a kind
Response towards the word bear.
of animal)
Respondent 1: dislike, strong, active.
Respondent 2: dislike, weak, passive.
1)Steps to forming the semantic
differential scale.
a)Generating ideas.
-) The development of the scale begins by compiling items,
i.e, collecting as many word pairs with opposite meanings
as possible that can represent the concept.

-) Ideas can be obtained by referring to past research or be


based on theories regarding the concept.

-) The researcher cannot make up his own word pairs to


represent the concept because without sound theories or
evidence, reliability of measurement will be low.
b) Evaluating the suitability of items.
- A panel of experts consisting of individuals with in-depth
knowledge and experience of the concept will evaluate
the items in three dimensions, which are the activity
dimension (high or low activity), evaluation dimension
(compatibility with concept) and potential dimension (low
or high capability).

c) Placing items in the semantic


differential scale.
- Selected items will be placed into the semantic
differential scale.

- Scales used will be 1-5, 1-7 or 1-9 (see figure 16.21 for
example of an item of the semantic differential scale,
page 336)
2) Analysis of semantic differential scale
data.
- The semantic differential scale is given the code -2 to
+2for scale 1 to 5, code -3 to +3 for scale 1 to 7 and
code -4 to +4 for scale 1 to 9.

- It is an ordinal scale and data analysis is similar to that of


the Likert scale. An analysis of the differences between
the respondent groups can be conducted using non-
parametric tests such as the Mann-Whitney U test and
the Kruskall-Wallis H test.

- Correlation analysis can be conducted using the


spearman rho test, whereas descriptive analysis can be
done using frequency and percentage.
Q&A
1.What is the differences
between Index and Scale?
2.List 3 standard
measurement scale used in
social science are and
educational research.
3.What are the advantage of
using Likert Scale?
Thank
You

S-ar putea să vă placă și