You are on page 1of 5

ED 502: EDUCATIONAL STATISTICS

TESTS OF RELATIONSHIP

STATISTICAL FUNCTION USE LEVEL/SCALE OF SAMPLE DATA SAMPLE PROBLEM


TREATMENT MEASUREMENT

It is used to determine if there is Interval / Ratio Scores, Grades, Ratings Is there a positive correlation between
Pearson - r
a correlation or relationship Reading Comprehension and
between two variables of the Ex. Personality of PNU graduate students?
interval or ratio type. Reading Comprehension
Personality Is there a significant relationship
between the Student Academic
Achievement in Math II and
Instructional Level among PNU
undergraduate students?

It is used to determine if there is Ordinal Ranks, Non-numeric Is there a correlation between


Spearman Rank-
a correlation or relationship scales Motivation and Self-Concept?
Order Correlation / between two variables of the
Spearman Rho ordinal type. Ex. What type of affective need is
Satisfaction addressed by the type of management
* nonparametric statistics Happiness of emotions?
(distribution free statistics) Motivation
Self-concept
* product-moment correlation Discomfort

It is used to determine if there is Nominal Categorical / 50 Male high school students and 50 Female
Chi-Square Test of high school students were asked about their
a correlation or relationship Classificatory Variables
Independence between two variables of the color preferences. Among the Male students,
nominal type. Ex. 20 preferred light colors, 15 preferred bright
and 15 preferred dark. Of the 50 Female
Sex
students, 15 preferred light colors, 15
Drink Preference
preferred bright and 20 preferred dark colors.
Color Preference
Is sex related to color preferences?
TESTS OF DIFFERENCE

STATISTICAL FUNCTION USE LEVEL/SCALE OF SAMPLE DATA SAMPLE PROBLEM


TREATMENT MEASUREMENT

It is used to determine if the Interval / Ratio 2 groups single interval The given population obtained a mean of 120
One Population and the standard deviation is 20. A sample of
given sample mean was drawn variable
Z - Test from the population with no 100 was drawn from the population and it
parameters or if a given group Ex. obtained a mean of 119.
represent the population. Group of Teachers
(sample mean) from the Test the hypothesis that the sample
* random / purposive sampling University (population) mean of 119 was drawn from the
Group of Teenagers population whose mean is 120.
(sample mean) from the
Metro Manila
(population)

It is used to determine if there is Nominal Frequencies converted to 50 Male graduate students and 50 Female
Z - Test of graduate students were asked whether they
a significant difference between Proportions
Independent two independent / different (2 groups 2 responses) are in favor or against same sex marriage.
groups or situations that call Among Male respondents, 25 are in favor and
Proportions 25 are against. Among Female respondents,
for two types of responses Ex.
20 are in favor and 30 are against.
(dichotomous). Yes and No
Agree and Disagree
Is there a significant difference between
* include issues in which you are In Favor and Against
Male and Female graduates who are
only asked whether you agree
in favor of same sex marriage?
or disagree
90 Ph.D candidates in Educ. Mgt. and 80
Ph.D candidates in Math respond to an item
concerning the usefulness of Foreign
Language requirement for the Ph.D. 30
Educational Mgt. candidates and 55 Math
candidates agree with the statement.

Is there a significant difference between


the proportions in each group that
responded In Favor?
It is used to determine if there is Nominal Pair of Observations Is there a significant difference in the
Z - Test of
a significant difference between responses of the 50 voters before and
Dependent pair of observations from a Ex. after the Meeting de Avance?
Proportions single group. Before and After
In Favor and Against Consider the test items A and B. In a sample
* response to a question given in Passed and Failed of 100 children, 30 pass item A fail item B,
two different occasions whereas 20 fail item and pass item B.

Are the proportions of children


passing the two items significantly
different from each other?

It is used to determine if there is Interval / Ratio 2 groups Control and Is there a significant difference between
T - Test of
a difference between two Experimental Group the experimental and control group in
Independent / groups using dependent terms of achievement in Statistics?
Uncorrelated independent variables. Ex.
Method of Teaching A survey was conducted on attitude toward
Means mentally impaired children. A random
* comparison of two groups Achievement
IQ Scores sample of teachers and administrators were
selected and asked to respond to an attitude
toward mentally impaired children scale.

Is there a significant difference between


teachers and school administrators in
terms of attitude toward mentally
impaired children?

It is used to determine if there is Interval / Ratio 2 groups Pre-test and Is there a significant difference between
T - Test of
a significant difference between Post-test the Pre-test and Post-test in Statistics
Dependent / two sets or two groups of of the Experimental Group?
Correlated Means correlated scores or measures. Ex.
Scores Is there a difference in the Pre-test and
Post-test scores in the Math
Achievement test of Grade IV pupils
when exposed to traditional method?
calculator-integrated method?
It is used to determine if there is Nominal Frequencies, Scale, 100 students were asked about their opinion
Chi-Square Test of about Pacquiaos candidacy. 23 strongly
difference between the Opinion, Response
Goodness of Fit observed (hypothetical / agreed, 21 agreed, 5 expressed no opinion, 25
disagreed and 26 strongly disagreed.
theoretical) distribution and Ex.
expected (predetermined) Responses
distribution. Strongly Agree (SA) Is there a significant difference between
Agree (A) the observed distribution and
No Opinion (NO) expected distribution of the students
Strongly Disagree (SD) responses on the said issue?
Disagree (D)
Are the observed frequencies the same
as an expected set of frequencies? (Ex.
Are the number of workplace accidents
the same for each hour of the day?)

It is used to determine if there is Interval / Ratio 2 or more groups Do the three groups of graduate
One-Way Analysis
a significant difference between students (discovery approach, distance
of Variance two or more groups in terms of Ex. learning, instructional TV) differ
(ANOVA I) means (achievement, weight, IQ). Method of Teaching: significantly in terms of achievement in
I Discovery Approach Statistics?
* extension of T-test / inspects II Distance-Learning
an independent variable III Instructional-TV

It is used to determine if there is Interval / Ratio 2 or more groups A researcher was interested in whether an
Two-Way Analysis individual's interest in politics was influenced
an interaction / combined
of Variance effect between or among two or Ex. by their level of education and gender.
Therefore, the dependent variable was
(ANOVA II) more independent variables Effect of Method of
"interest in politics", and the two independent
(ex. method of teaching, medium Teaching and Medium
variables were "gender" and "level of
of instruction) to the dependent of Instruction to
education".
variable (ex. achievement test). Achievement
Is there an interaction between
* can answer 3 questions education level and gender / was the
effect of level of education on interest in
politics different for males and females?
LEVELS OF MEASUREMENT

NOMINAL ORDINAL INTERVAL RATIO

Chi-Square Test of Independence Spearman Rank-Order Pearson - r


STATISTICAL
TREATMENTS Z - Test of Independent Proportions Correlation / Spearman Rho One Population Z - Test

TO BE USED Z - Test of Dependent Proportions T - Test of Independent / Uncorrelated Means


Chi-Square Test of Goodness of Fit T - Test of Dependent / Correlated Means
One-Way Analysis of Variance (ANOVA I)
Two-Way Analysis of Variance (ANOVA II)

STATISTICS / proportions Median Mean


MEASURES percentage Quartile Deviation Standard Deviation

PRE-TEST POST-TEST CONTROL GROUP DESIGN: GUIDE IN CHOOSING STATISTICAL TREATMENT:

1) Determine the type of research question to be answered by the statistical analysis.


a. the degree of relationship or dependence among variables (H0 = there is no relationship or dependence,
EXPERIMENTAL GROUP CONTROL GROUP
and the statistical test answers the question as to whether any relationship or dependence found is
sufficiently different from zero that it can be considered statistically significant.)
Pre-test Post-test Pre-test Post-test b. the significance of group differences (H0 = there is no difference between groups. The statistical test
answers the question as to whether an observed difference is probably due just to random factors, or is
large enough to be considered statistically significant and due to the treatment factor.)
T - Test of Dependent T - Test of Dependent
T - Test of Independent Means Means
2) Determine the nature(s) of the variables under discussion, and whether they meet the assumptions of a
Means
particular test (e.g. the data are normally distributed).

3) Types/levels of data:
LEVEL OF SIGNIFICANCE NOMINAL unordered categories; numbers simply express identity/ labeling purposes (e.g. religion;
country of birth; sex; etc.)
.05 you have 5 chances (Type I Error) out of 100 that you are wrong with your ORDINAL ordered categories; numbers express ranks (e.g. level of agreement on an opinion survey;
decision and 95% that youre correct; rejecting the null hypothesis when it should
proficiency level at a martial art; scale used in determining the hardness of a mineral, academic ranks)
be accepted; has smaller Region of Acceptance and bigger Region for Rejection (
for nominal and ordinal data, what is usually recorded is the number of occurrences of a particular
any computed Z-value or absolute value 1.96, reject Ho )
result (e.g. number of Christians, number of Buddhists etc. but these numbers are not the values of the
.01 you have 1 chance (Type II Error) of being wrong and 99% that youre variable. In this case, variable = religion, values = Christian, Buddhist, and the numbers are the number
correct; accepting the null hypothesis when it should be rejected; has bigger Region of occurrences of a particular value.)
of Acceptance and smaller Region for Rejection ( any computed Z-value or absolute INTERVAL ordinal + distance between values is of constant size; has equal intervals that is distances
value 2.58, reject Ho) between points are equal starting an arbitrary zero (e.g. temperature; score)
RATIO interval + (i) there is a meaningful zero and (ii) the ratio between two numbers is meaningful;
DECISION INTERPRETATION OF HYPOTHESIS zero is absolute; there is absolute value of the variable (e.g. height; weight, distance, number of children)
Significant REJECT use Alternate Hypothesis ratio and interval data can be either discrete (i.e. there are gaps between values, e.g. number of
children) or continuous (i.e. there are no gaps between values (e.g. weight, height).
Not Significant ACCEPT use Null Hypothesis