### Present Remotely

Send the link below via email or IM

• Invited audience members will follow you as you navigate and present
• People invited to a presentation do not need a Prezi account
• This link expires 10 minutes after you close the presentation

Do you really want to delete this prezi?

Neither you, nor the coeditors you shared it with will be able to recover it again.

# Statistics Presentation

No description
by

## Seung-yon Kang

on 3 April 2014

Report abuse

#### Transcript of Statistics Presentation

RESEARCH DESIGN
English Preference Tendencies and English Learning Anxiety in Primary School Children
VALIDITY
Reliability
Variables
Validity
Variables
Independent and Dependent Variables
The independent variable(IV)
is that we believe may
"cause"
the results

The dependent variable(DV)

is that we measure to see the

effects
due to the independent variable.

Moderator Variables
- A type of independent variable.
- It may not be the main focus of the study,
but may modify the relationship
between the IV and the DV.
- It may result in an interaction between
an IV and other variables.
Intervening Variables
(cf. moderator variables)

They are not included in an original study
- They cannot be
identified in a precise way
.
- The researcher has not considered the
possibility of their effect

Control Variables
- A variable that is held constant across groups
in order to eliminate the effect of that variable
on the outcome of the study
- Whenever possible, researchers need to
identify possible factors that might affect
the result and control for them in some way
independent vs. dependent
moderator
intervening
control
THANK YOU !
Research Questions
Method
Materials
Data analysis
Conclusion
Research Questions
Q1
. To investigate the correlation between
English Preference Tendencies (EPT) and English Learning Anxiety (ELA)
in Korean primary school children

Q2
. To identify the factor influencing children’s EPT
(learner variables: grades, grade, starting age, learning experience outside public school classrooms, interaction with family members, experience abroad, and the amount of time invested in English learning)

Method
1)
Design
: a pilot study + final survey
2)

Procedure
Interview
(2 English teachers)
Classroom observation
Develop questionnaire items
25 items: individual learner factors (5), EPT (10), ELA (10)
Pilot survey
Final survey
3)
Participants
- 3rd ~ 6th grades in a Korean primary school
- 3 classes from each grade were randomly chosen (from 1 school)
- Total number: 400 (Each class: 33 students)

*9 students did not complete the questionnaire.
Only 391 were considered valid data.

Materials
Questionnaire (3 parts)

1) Individual Learner Factors Related with English Learning
2) English Preference Tendencies Measurement for Children (
EPT
MC)
3) English Learning Anxiety Measurement for Children (
ELA
MC)

Data Analysis
- To measure how two variables (EPT & ELA of children) are related

Correlation analysis (ANOVA)

- To examine the effect of learner variables on EPT in children

One-way ANOVA

A post-hoc test
- To see the differences between the subjects

- For the effect of grade & gender on EPT & ELA

MANOVA
Conclusion
- Interacting with family members for learning English showed
significantly higher EPT (vs, not doing home study at all)
Content
validity
Face validity
Construct validity
Criterion-related
validity
Predictive
validity
Internal validity
External
validity
Content Validity

Representativeness
of our measurement

e.g. study acquisition of relative clauses
"The boy
who
is running is my friend"
(only one type of relative clause)
Face Validity

The
familiarity of our instrument

How easy it is to
convince others
that there is content validity.

- Research on elementary school students
Construct Validity
How well the research
defines the construct
(element) of interest.
An essential topic in SLA research : many of the
variables are not easily defined
.

(e.g. aptitude, exposure to input)
To enhance:
Multiple estimates
of a construct

e.g. - Amount of input
- KCSAT. Listening Q.17

(reading is not a construct of listening ability)

Criterion-related Validity
tests used in a research study

other well-established tests
Predictive Validity
Whether the current research can
predict performance
on other situation.

e.g.
-SAT score : college grade (GPA)
-Middle school 1st grade. Occupation aptitude test
20 years later, Actual job?

Internal Validity
External Validity
Internal Validity
Are the differences found for the dependent variable
directly related
to the independent variable?
A researcher must
control
for all
other possible factors
that might affect the results.

e.g. learners respond by pressing a button(T,Y)
on the computer.
(left-handed people might react faster)

External Validity
The findings of the study are relevant not only to the research population, but also to the
wider population
of language learners.
Validity
Internal
External
Activity 1
Variables
- Characteristics that vary

- Features or qualities that change.
e.g.
- Many language programs : their own internal tests
- A test that I developed for my research =
TOEIC, TOFLE score
e.g. The participants chosen for any study
form a research population.
-Participant
Characteristics

-Participant
Mortality

-Participant
Inattention and Attitude
-Participant
Maturation

-Data Collection:
Location and Collector

-Instrumentation
& Test Effects
language background, language learning experience, proficiency level
the loss of participants during the research
Pre test – treatment – post test
Sampling
Representativeness & Generalizability
Collecting Biodata Infomation
[positive effect]
-
Hawthorne effect
: participants know that they are part of an experiment
-
Halo effect
they think are expected.
[negative effect]
-
fatigue and boredom
-
general inattentiveness
: participants stated that his results from the two
sessions differed.
e.g. participant was thinking about final exam- not fully concentrate
Usually in longitudinal studies, when children is involved
(they grow old)
Location
: Physical setting
one group- noisy and uncomfortable class, the other group is not.
Collector
English interviewer vs. Korean interviewer
-Consider the
representativeness

of the sample
-Our participants need to be
chosen
randomly
from the population.

Sampling
-
Equivalence between pre- and posttests

Pre and post test needs to be equal level
-
Giving the goal of the study away

participants already know the goal >> affect the result

e.g. pre test >> alert participants

1st solutions
: conducting the pretest a few weeks before the study
(students don’t know the relationship)

2nd solutions
: mixing research-related questions in a much larger test
-
Test instructions and questions
appropriate to the developmental level
Nonrandom Sampling
Random Sampling
-The (instrument)
consistency

of a study

e.g. same student took the same listening test twice
1st test: 80 / 2nd test:40 = not reliable test
Rater Reliability
Instrument Reliability
Reliabilityㄱㄷㄷ
the characteristic of rater
Instrument Reliability
To ensure our
instrument reliable

[ 3 types of reliability testing ]

Rater Reliability
Inter
rater Reliability
Intra
rater Reliability
Two or more raters judge the same set of data in the same way.
Whether the researcher judge the data same ways at different times
Test-Retest
Equivalence of Forms
Internal consistency
Test-Retest
- Giving the

same test

to the

same group

at

two different times
- Determine the appropriate

time interval

between the tests.

Equivalence of Forms
Two version

of a test are administered to the

same individuals

and see the relationship between two scores.
e.g. Two version of a test -KCSAT (odd, even number set)

Internal consistency
It is not always possible to administer tests twice
so we use

statistical methods

to determine reliability

Split-half procedure
: divide one test into two parts.
Kuder-Richardson 20 and 21
: using the number of items, the mean, and the standard deviation
(more details: chapter 9)
Cronbach’s
Simple
random
sampling
Random Sampling
Operationalization
-When it is difficult to measure
variables
directly,

working definitions
of variables
-An operational definition
allows researchers
to work with
the variables. (allow measurement)
Measuring Variables: Scales of Measurement
Nominal scales
researchers
categorize variables
into two or more groups.
e.g. assign male and female group.

Ordinal scale
:

ordering is implied.
no implication of an equal distance between each rank order.
e.g. student test scores are often ordered from best to worst.

Interval scale:
the order like ordinal scale
but it also reflects the
distance

e.g. test scores
the distance between 70 - 80
same
the distance between 80 - 90
Stratified random Sampling
Cluster
random
sampling
How to measure?
Activity 2
Simple random sampling
Every member of a population has an
equal chance of being selected

- Researchers have to select representative participants from the entire population randomly.

Stratified random Sampling
The
proportions
of the
subgroups
are first determined, and then participants are randomly selected.
Simple random sampling
Representativeness and Generalizability
To
generalize
the results of a particular study,
representativeness of the sample
is important.

e.g. a study conducted in a university setting
private language school setting
Collecting Biodata Information
Include
sufficient information
!
so that the reader can determine whether the results are generalizable to a new context.
Nonrandom
Sampling
Systematic sampling
Convenience sampling
Purposive
sampling
Systemic sampling
The choice of every
n
th individual in a population list
Convenience sampling
The selection of individuals who are available for study (weak point:
not representative)
e.g. surveying passersby in a mall, friends, or colleagues
sample biodata form
Purposive sampling
Researchers
intentionally select
individuals
(weak point:
not representative)
Skills course: organizing a new group
the seating was random
Cluster random Sampling
The selection of
groups
rather than individuals as the objects of study
correlation analysis
English
Preference
Tendencies (EPT)
English
Learning
Anxiety (ELA)
statistically significant
Grade, Starting Age of learning English, learning experience outside public school settings, interaction with family members
The effect of Learner Factors on EPT
significantly influenced the overall EPT
Learner’s gender, the amount of time invested in English learning, and experience abroad
did not show any significant effect on EPT
Students’ overall EPT is becoming lower and their ELA higher