View
220
Download
0
Category
Preview:
Citation preview
8/11/2019 Els Sfr14a Hdout Genrl
1/23
0
Handout (to be improved whilst sessions in progress)
Statistics for Research
The Graduate Program of English Studies
Sanata Dharma University
Feb 2014
8/11/2019 Els Sfr14a Hdout Genrl
2/23
1
PREFACE
Why learn statistics for research in English language
studies? What for?
You are going to become an expert in English language
studies. You will become able to manage research to solve English
studies problems scientifically. Since all research involves logic and
observation, their coherence (validity) remains crucial for every
research project. Statistical concepts help clarify this, regardless
whether your research is with or without statistics
What do you compose this learning portfolio for?
To build your own perspectives of statistics for research, toenable you to describe and interpret a research validity, and improve
it.
What do you do?
You describe every issue briefly, state what it is for (the
goal), how to achieve the goal, and what is needed to process to
achieve the goal. All needs cross-checking with recent literature.
Which is most viable?
It is not the idea that every one of you shall elaborate every
issue, but only the most viable ones, given the allocated time and
your initial knowledge. The idea is that we make the most of the
time and other facilities at our disposal. We will at least spend
ample time on Part1, Part 2 Chapter IX: Validation of ResearchDiscovery, and electronic data processing (group project).These
having been done, it is reasonable to expect that you will be able tovalidate your own research.
8/11/2019 Els Sfr14a Hdout Genrl
3/23
2
TABLE OF CONTENTS
PREFACE
TABLE OF CONTENTSPart 1 Research and Statistics
I. RESEARCHA. MEANING
B. FUNCTIONS AND TYPES
II. STATISTICS
A. MEANING
B. FUNCTIONS AND TYPESC. CONCEPT, OBSERVATION, AND EXPERIENCE
D. OBJECT, ATTRIBUTE, PHENOMENON, VARIABLE
E. MEASUREMENTF. DATA
G. INTERPRETATION
Part 2 Statistical Techniques
III. DESCRIPTIVE STATISTICS
A. MEANING
B. DATAB. MEASURES OF CENTRAL TENDENCY
C. MEASURES OF DISPERSION
D. CORRELATION
IV. INFERENTIAL STATISTICS
A. PARAMETRIC
8/11/2019 Els Sfr14a Hdout Genrl
4/23
3
B. NON-PARAMETRIC
V. STATISTICS FOR SURVEY RESEARCH
A. SIMPLE CORRELATION
B. MULTIPLE CORRELATION
VI. STATISTICS FOR EXPERIMENTAL RESEARCH
A. SIMPLE COMPARISON
B. MULTIPLE COMPARISON
VII. STATISTICS FOR OTHER RESEARCH
A. EXPLORATORY RESEARCHB. RESEARCH AND DEVELOPMENTC. CONTENT ANALYSIS
VIII. SELECTION OF A STATISTICAL TECHNIQUE
A. KNOWLEDGE PRE-REQUISITES
B. PROCEDURE
IX. VALIDATION OF RESEARCH DISCOVERYA. VALIDITY
B. HYPOTHESIS FORMATION
C. EMPIRICAL VERIFICATION (HYPOTHESIS TESTING)
D. INTERPRETATION (OF DATA PROCESSING RESULTS)
8/11/2019 Els Sfr14a Hdout Genrl
5/23
4
Part 1 Research and Statistics
I. RESEARCH
A. MEANING
A systematic investigation to discover the truth, the
scientific truth
Systematic investigation: think, plan, do
The scientific truth: logical and empirical/experiential
Logical truth: deductive, inductive
Empirical: about what you see using sensory organsConfirmatory (quantitative): deductive-inductive
Exploratory (qualitative): inductive-deductive
Experiential: see sensuously (statistically irrelevant)
B. FUNCTIONS AND TYPES
EDUCATIONSystematic learning to improve life quality, empirical and
experiential
Empirical life-quality becoming more efficient and
effective (e.g. learners higher speaking ability, teachers
less preparation time, texts becoming more motivating)
Experiential life quality (statistically irrelevant)::
promotion of equity leading to self-actualization, personal
and social (= becoming excellent in ones own right)
C. ENGLISH
ELF, WE, IE, GE rather than EFL, ESOL, ESL
Focus: regional English; how to deploy it rather than
master it
8/11/2019 Els Sfr14a Hdout Genrl
6/23
5
English education research is necessarily pragmatic. It is
not only to discover the scientific truth, but to imply a
contribution to life-quality improvement, empiric orexperiential.
ENGLISH EDUCATION RESEARCH VALIDITYValidity = coherence between what is meant to investigate with
what actually is investigated
What is investigated = the scientific truth of English education =
a theory
A theory = a logical (and empirical) relation of concepts
Internal validity = degree of coherence between concepts and
data
External validity = the discovered theory and real life
(statistically: sampling and population)
8/11/2019 Els Sfr14a Hdout Genrl
7/23
6
II. THE STRUCTURE AND FUNCTIONS OF
STATISTICS
A. RESEARCH VALIDITY AND STATISTICS
B. CONFIRMATORY RESEARCH(deductive-inductive; hypothesis testing)
a. Hypothesis of association
Ho: no associationH1: there is (only; positive; negative)
b. Hypothesis of difference
Ho: no difference
H1: there is; larger; smaller
C. EXPLORATORY RESEARCH
(theory generating; inductive-deductive)1. Non-inferential
2. Inferential
8/11/2019 Els Sfr14a Hdout Genrl
8/23
7
II. STATISTICS
A. MEANINGTraditionally (mechanistic): statistics is the science of conductingstudies to collect, organize, summarize, analyze, and draw
conclusions from data (Bluman, 2000?)
Research related (conceptual): to ensure research validity,internal and external, i.e. by keeping consistent relation between
logic and observation, notably the essential components of scientific
research
1. Statistics
i. Observation
ii. Object, attribute, phenomenoniii. Data
iv. Organization
v. Process
vi. Interpretation
2. A statistic
i. indexii. analysis technique
iii. vs. parameteriv. ...
8/11/2019 Els Sfr14a Hdout Genrl
9/23
8
1. Statistics
i. Observation
ii. Object, attribute, phenomenon
iii. Data
iv. Organizationv. Process
vi. Interpretation
2. A statistic
i. index
ii. analysis technique
iii. vs. parameter
iv. ...
B. FUNCTIONS AND TYPES
C. CONCEPT, OBSERVATION, AND EXPERIENCE
1. Meaning
2. Procedure
i. Elicitation
ii. Measurementiii. Record
D. OBJECT, ATTRIBUTE, PHENOMENON, VARIABLE
E. MEASUREMENT
F. DATA1. Meaning
recorded result of observation
to represent a world reality2. Observation
3. Recording
4. Variable
5.
8/11/2019 Els Sfr14a Hdout Genrl
10/23
9
G. INTERPRETATION
8/11/2019 Els Sfr14a Hdout Genrl
11/23
10
Part 2 Statistical Techniques
III. DESCRIPTIVE STATISTICS
A. MEANING
B. DATA
B. MEASURES OF CENTRAL TENDENCY
C. MEASURES OF DISPERSION
D. CORRELATION
8/11/2019 Els Sfr14a Hdout Genrl
12/23
11
IV. INFERENTIAL STATISTICS
A. PARAMETRIC
1. Correlational
i. Pearson r
ii. Regression
2. Comparative
the t-test
ANOVA
B. NON-PARAMETRICRequirements violated, or, nominal and ordinal data
1. Chi-square
2. Mann Whitney U-Test
3. Kruskal Wallis H Test
4. Wicoxon Test
5. Friedman Test
8/11/2019 Els Sfr14a Hdout Genrl
13/23
12
V. STATISTICS FOR SURVEY RESEARCH
A. SIMPLE CORRELATION
B. MULTIPLE CORRELATION
8/11/2019 Els Sfr14a Hdout Genrl
14/23
13
VI. STATISTICS FOR EXPERIMENTAL
RESEARCH
A. SIMPLE COMPARISON
B. MULTIPLE COMPARISON
8/11/2019 Els Sfr14a Hdout Genrl
15/23
14
VII. STATISTICS FOR OTHER RESEARCH
A. EXPLORATORY RESEARCH
1. Descriptive (non-inferential): descriptive statistics
2. Inferential: Factor Analysis; SEM
B. RESEARCH AND DEVELOPMENT
(Validation of conceptual and iconic model)
1. Non-inferential: descriptive statistics
2. Inferential: Factor Analysis; SEM
C. CONTENT ANALYSIS
8/11/2019 Els Sfr14a Hdout Genrl
16/23
15
VIII. SELECTION OF A STATISTICAL
TECHNIQUE
A. KNOWLEDGE PRE-REQUISITES
1. Variable
a. one: descriptive
b. two
1) descriptive: descriptive
2) inferential: inferentialc. more than two: multiple
2. Hypothesis (Variable relation)
a. Association
1) simple
2) multiple
b. Difference
1) simple2) multiple
c. Causal: independent-dependent; non-causal
3. Groups (within a variable)
a. Two or more
b. Independent (between subjects) or correlated (within subjects)
4. Data
a. Discrete or continuous
b. Data scales: nominal, ordinal, interval, ratioc. Standard scoring
c. Normal distribution
5. Parameter
a. Statistics and Parameters
8/11/2019 Els Sfr14a Hdout Genrl
17/23
16
b. Definite and non-definite (hypothesis testing)
B. PROCEDURE
1. The nature of researcha. Descriptive, non-inferential
b. Inferential
2. Decide your hypothesis: association or difference
3. Data scale: nominal/ordinal, non-parametric; interval/ratio:
probably parametric
a. If normally distributed, linear, and homogeneous: parametric
b. If (a) strongly violated: non-parametric4. Measurement: correlated or independent5. Hypothesis of association
a. Data scale
b. Without prediction:
c. With prediction:
d.
6. Hypothesis of difference
a. Data scaleb. Number of independent variablesc. Correlated data
d.
7.
Table 1.Statistical Technique-Selection: Descriptive, Non-
Inferential Research*)
Measure
Data
Nominal Ordinal Interval &
Ratio
Central
tendency
mode median average
Dispersion label range standard
deviation;
8/11/2019 Els Sfr14a Hdout Genrl
18/23
17
variance
Location label percentile z-score; T-
score
Association: point biserial;
biserial
Spearman rho Pearson r
*)Povisional. Please cross-check with references for correction and
completion
Table 2.Statistical Technique-Selection: Inferential Research*)
Hypothesis Correlated IndependentParamtrc Non-Par Paramtric Non-Par
ASSOCIATION
2 Var
predict
Pearson r
Simple
Reg
Spearma
n rho;Chi-
square
- -
> 2 Var Multiple
Reg
Chi-
square
- -
DIFFERENCE
1IV
2 groups
> 2 groups
t-test (cor)
Anova
Wilcoxon
Test
FriedmanTest
t-test (ind)
Anova (1-way)
Chi-
square;
Mann
Whitney
KuskallWallis
2IV Anova (2-
way)
Anova (2-
way)
*)Povisional. Please cross-check with references for correction andcompletion
8/11/2019 Els Sfr14a Hdout Genrl
19/23
18
IX. VALIDATION OF RESEARCH
DISCOVERY
A. VALIDITY
Validity: coherence of what you do with what you have intended to
do
Research-project validity: coherence of the research project with theparticipation in improving life quality
Validity of scientific (logical positivistic) research-project: whetheryou discover a theory (which is verified or generated)
Why theory? Because theory explains, predict, and control better, so
you can perform better.
A theory: a relation of variables (logical and empirical)
A variable, and variable relation, is a construct (= concept). It is
abstract. It exists in your mind only.
There are a countless number of variables in any academic
discipline, English education being no excempt. To make sense,
every one of them needs defining, discernedlyfrom others. This iswhat is called construct validity. Reading ability, for example, has to
be reading ability, and not other variables like grammar ability,vocabulary, or general intelligence. Construct validity is the most
important validity.
Again, your research discovery is a theory. In general, you will
verify (or test) a theory, mainly for two reasons. First, it is not easy
to identify an English education issue the theory of which is still
8/11/2019 Els Sfr14a Hdout Genrl
20/23
19
scarce. Secondly, theory generation needs a wide and large data
size, which is unlikely accessible withion your time allocation.
A theory, as stated before, is a relation variables. Understandably,
for a relation to exist there must at least two variables (Please note,
that the relation itself is also a variable). A theory is also called athesis. It is a construct (= concept) which is both logically true and
empirically true. When it is logically true only, it is called a
hypothesis. Coincidentally your hypothesis is also the answer
(logical only) of your research question.
Validity of research discovery
Logically true variablesLogically true relationsLogically true variable operationalization
Logically true relation operationalization
Data representativeness
Coherent (methodologically correct) data processing
Coherent (methodologically correct) interpretation of data
processing result
Internal validity = validity of research discovery
External validity = applicability in a similar context other than the
research setting
B. HYPOTHESIS FORMATION
1. Clarification of variables and relevant concepts, including
working definitions of variables and variable relations2. Statement of universal theory
3. Clarification of local context
4. Logical relation of 2 and 3 = conceptual hypothesis5. Statistical hypothesis
C. EMPIRICAL VERIFICATION (HYPOTHESIS TESTING)
1. Variable operationalization
a. Construct validation (blueprint)
8/11/2019 Els Sfr14a Hdout Genrl
21/23
20
b. Data: nature, types
c. Data gathering instruments (including their validation)
2. Data presentation
3. Data processing: computation, conclusion, decision
D. INTERPRETATION (OF DATA PROCESSING RESULTS)
1. Statistical: descriptive statistics, inferential statistic values,
significance level, hypothesis acception/rejection
2. Scientific: conceptual hypothesis
3. Recommendation
8/11/2019 Els Sfr14a Hdout Genrl
22/23
21
SCORING
No Statistics &
Research
Competence
Know-ledge
Skill
Attitude
Con-science
Compas-sion
1 2 3 4
1 Research x
2 Statistics
3 Statistics &
Research
x
4 DescriptiveStatistics
5 Inferential
Statistics
11 Research knowledge
12 Research skill (NA)21
2231
32
41
42
51
52
12345.3 Conscience
Primacy of genuine progress rather than formality (contribution
to progress: self-, group, and class; absence of plagiarism and
cut-and-paste work).
12345.4 Compassion
8/11/2019 Els Sfr14a Hdout Genrl
23/23
22
Show the implications of statistics and statistical research to the
promotion of human dignity and suggest how to offset
potentially negative effects
Recommended