1.
Using the data for our class exercise test whether performance in English, performance in
Math, type of school graduated from and sex are predictors of performance in Science.
State the prediction model. How much variance is accounted for by each of the predictor
variables? Assume all the assumptions for parametric test are met.
a) State the null and alternative hypotheses.
Ho: Performance in English, performance in Math, type of school graduated
from, and sex are not predictors of performance in science.
Ha: At least one of the independent variables (Performance in English,
performance in Math, type of school graduated from, and sex) is/are
predictors of performance in science.
b) Set alpha level.
𝛼 = .05, 2-tailed.
c) Statistical Tool
Multiple Regression Analysis
d) Result
Descriptive Statistics
Mean Std. Deviation N
Test Score in Science 82.00 6.787 35
Test Score in English 83.91 6.934 35
Test Score in Math 82.66 6.188 35
Sex 1.51 .507 35
Type of Prinary School
1.40 .497 35
Graduated From
Correlations
Test Test Test Type of Primary
Score in Score in Score in School
Science English Math Sex Graduated From
Pearson Test Score in Science 1.000 .910 .897 .085 .262
Correlation
Test Score in English .910 1.000 .872 .063 .369
Test Score in Math .897 .872 1.000 .011 .361
Sex .085 .063 .011 1.000 -.023
Type of Primary School Graduated From .262 .369 .361 -.023 1.000
Sig. (1- Test Score in Science . .000 .000 .313 .065
tailed) Test Score in English .000 . .000 .359 .015
Test Score in Math .000 .000 . .475 .016
Sex .313 .359 .475 . .447
Type of Primary School Graduated From .065 .015 .016 .447 .
N Test Score in Science 35 35 35 35 35
Test Score in English 35 35 35 35 35
Test Score in Math 35 35 35 35 35
Sex 35 35 35 35 35
Type of Primary School Graduated From 35 35 35 35 35
Variables Entered/Removeda
Variables Variables
Model Entered Removed Method
1 Type of Prinary
School
Graduated
From, Sex, Test . Enter
Score in Math,
Test Score in
Englishb
a. Dependent Variable: Test Score in Science
b. All requested variables entered.
Model Summaryb
Change Statistics
Mode R Adjusted R Std. Error of R Square F Sig. F
l R Square Square the Estimate Change Change df1 df2 Change
1 .941a .885 .869 2.455 .885 57.458 4 30 .000
a. Predictors: (Constant), Type of Prinary School Graduated From, Sex, Test Score in Math, Test Score in English
b. Dependent Variable: Test Score in Science
ANOVAa
Sum of Mean
Model Squares df Square F Sig.
1 Regressio
1385.190 4 346.298 57.458 .000b
n
Residual 180.810 30 6.027
Total 1566.000 34
a. Dependent Variable: Test Score in Science
b. Predictors: (Constant), Type of Prinary School Graduated From, Sex,
Test Score in Math, Test Score in English
Coefficientsa
Unstandardized Standardized
Coefficients Coefficients Correlations
Model B Std. Error Beta t Sig. Zero-order Partial Part
1 (Constant) -3.283 5.842 -.562 .578
Test Score in English .535 .126 .546 4.256 .000 .910 .614 .264
Test Score in Math .502 .140 .458 3.588 .001 .897 .548 .223
Sex .582 .836 .044 .697 .491 .085 .126 .043
Type of Prinary School
-1.426 .916 -.104 -1.557 .130 .262 -.273 -.097
Graduated From
a. Dependent Variable: Test Score in Science
2
STEPWISE METHOD
Variables Entered/Removeda
Variables Variables
Model Entered Removed Method
1 Test Score in Stepwise (Criteria: Probability-of-F-to-enter <= .050, Probability-of-F-to-
.
English remove >= .100).
2 Test Score in Stepwise (Criteria: Probability-of-F-to-enter <= .050, Probability-of-F-to-
.
Math remove >= .100).
a. Dependent Variable: Test Score in Science
Model Summaryc
Change Statistics
Mode R Adjusted R Std. Error of R Square F Sig. F
l R Square Square the Estimate Change Change df1 df2 Change
1 .910a .828 .823 2.856 .828 158.969 1 33 .000
2 .934b .873 .865 2.493 .045 11.303 1 32 .002
a. Predictors: (Constant), Test Score in English
b. Predictors: (Constant), Test Score in English, Test Score in Math
c. Dependent Variable: Test Score in Science
ANOVAa
Model Sum of Squares df Mean Square F Sig.
1 Regression 1296.801 1 1296.801 158.969 .000b
Residual 269.199 33 8.158
Total 1566.000 34
2 Regression 1367.066 2 683.533 109.951 .000c
Residual 198.934 32 6.217
Total 1566.000 34
a. Dependent Variable: Test Score in Science
b. Predictors: (Constant), Test Score in English
c. Predictors: (Constant), Test Score in English, Test Score in Math
Coefficientsa
Unstandardized Standardized
Coefficients Coefficients Correlations
Zero-
Model B Std. Error Beta t Sig. order Partial Part
1 (Constant) 7.261 5.947 1.221 .231
Test Score in 12.60
.891 .071 .910 .000 .910 .910 .910
English 8
2 (Constant) -.977 5.741 -.170 .866
Test Score in
.522 .126 .533 4.141 .000 .910 .591 .261
English
Test Score in
.474 .141 .433 3.362 .002 .897 .511 .212
Math
a. Dependent Variable: Test Score in Science
3
Excluded Variablesa
Collinearity
Partial Statistics
Model Beta In t Sig. Correlation Tolerance
1 Test Score in Math .433b 3.362 .002 .511 .240
Sex .028b .384 .703 .068 .996
Type of Prinary School
-.086b -1.105 .277 -.192 .864
Graduated From
2 Sex .048c .747 .461 .133 .988
Type of Prinary School
-.106c -1.601 .119 -.276 .857
Graduated From
a. Dependent Variable: Test Score in Science
b. Predictors in the Model: (Constant), Test Score in English
c. Predictors in the Model: (Constant), Test Score in English, Test Score in Math
e) Decision
Since, 𝑝 < 0.001 and is less than 𝛼 = 0.05, then we reject Ho.
f) Conclusion
A multiple regression was carried out to investigate whether performance in
English, performance in Math, type of school graduated from, and sex could significantly
predict the performance in science. The result of the regression indicated that the model
explained 87.3% of the variance and that the model was a significant predictor of science
performance, 𝐹(1,32) = 11.303, 𝑝 < .001. The science performance can be significantly
predicted by performance in English (𝑡 = 4.141, 𝑝 < .001), and performance in Math
(𝑡 = 3.362, 𝑝 = .002). However, Primary School Graduated from (𝑡 = −1.601, 𝑝 = .119)
and Sex (𝑡 = .747, 𝑝 = .461) are not significant predictors of science performance.
Furthermore, Science performance can be predicted using the model,
Science Performance = .522 (English Performance) + . .474 (Math Performance) – 977.
The result also indicates that 87.3% of the variance in the performance in
science of the students can be attributed by the performance in English and
performance in math.
Using the stepwise method, the results implied that 82.8% of the variance in
performance in science can be explained by the performance in English and 4.5% of the
variance in performance in science can be explained by the performance in Math.
However, sex and Type of school graduated from cannot account for the variability in
science scores because they are not significant predictors of the performance in
science.
4
2. The data below show the problem-solving skills and critical thinking skills of grade 10
students who have been taught through a lecture method (method 1) and through inquiry
approach (method 2). Are the means for the scores on problem solving skills and critical
thinking skills (or linear combinations of these test scores) the same or different for students
in the two study groups? Assume all the assumptions for parametric test are met.
Method 1 Method 2
Problem solving skills Critical thinking skills Problem solving skills Critical thinking skills
66 92 46 55
81 98 91 99
31 60 41 45
89 99 84 91
72 93 41 49
63 88 56 61
55 81 73 79
68 89 38 39
73 94 81 88
42 74 58 63
57 87 71 73
59 90 63 71
48 83 53 51
41 76 67 73
51 82 60 89
35 76 79 96
53 57 31 35
65 73 62 66
49 84 52 58
54 84 52 59
a) State the null and alternative hypotheses.
Ho: The mean scores on problem solving skills and critical thinking skills are
significantly the same between the two study groups of Grade 10 students.
Ha: The mean scores on problem solving skills and critical thinking skills are
significantly different between the two study groups of Grade 10 students.
b) Set alpha level.
𝛼 = .05, 2-tailed.
c) Statistical Tool
Multivariate Analysis of Variance
d) Result
Descriptive Statistics
Method Mean Std. Deviation N
Problem Solving Skills Lecture 57.60 14.922 20
Inquiry Approach 59.95 16.433 20
Total 58.78 15.539 40
Critical Thinking Skills Lecture 83.00 11.239 20
Inquiry Approach 67.00 18.935 20
Total 75.00 17.374 40
5
Multivariate Testsa
Hypothesis Partial Eta Noncent. Observed
Effect Value F df Error df Sig. Squared Parameter Powerc
Intercept Pillai's Trace .962 464.123b 2.000 37.000 .000 .962 928.246 1.000
Wilks' Lambda .038 464.123b 2.000 37.000 .000 .962 928.246 1.000
Hotelling's Trace 25.088 464.123b 2.000 37.000 .000 .962 928.246 1.000
Roy's Largest Root 25.088 464.123b 2.000 37.000 .000 .962 928.246 1.000
Method Pillai's Trace .578 25.306b 2.000 37.000 .000 .578 50.613 1.000
Wilks' Lambda .422 25.306b 2.000 37.000 .000 .578 50.613 1.000
Hotelling's Trace 1.368 25.306b 2.000 37.000 .000 .578 50.613 1.000
Roy's Largest Root 1.368 25.306b 2.000 37.000 .000 .578 50.613 1.000
a. Design: Intercept + Method
b. Exact statistic
e) Decision
There was statistically significant difference in the problem solving skills and critical
thinking skills of the students between the two study groups, 𝐹 (2, 37) = 25.306, 𝑝 <
.001; Wilk′s 𝚲 = 0.422, partial η2 = .578. Reject Ho.
f) Conclusion
The One-way Multivariate Analysis of Variance (One-Way MANOVA) was
carried out to determine whether the problem solving skills and critical thinking skills are
significantly different when they are grouped according to teaching methods. The result
reveals that the combined problem solving skills and critical thinking skills of the grade
10 students differ significantly when they are grouped according to teaching methods
(𝐹 (2, 37) = 25.306, 𝑝 < .001; Wilk′s 𝚲 = 0.422, partial η2 = .578). The result also
indicates that 57.8% of the variance in the problem solving skills and critical thinking
skills of the students can be attributed by the teaching methods used.
Furthermore, in the Test of Between-Subjects Effects, the critical thinking skills of
the students differ significantly with the type of teaching methods used, 𝐹 (1, 38) =
10.560, p = .002. The result infers that the critical thinking skills of the students exposed
to lecture and inquiry approach methods varies significantly. While, the problem solving
skills of the students do not differ significantly with the type of teaching methods used,
𝐹 (1, 38) = .224, p = .639. The result implies that the problem solving skills of the
students exposed to lecture and inquiry approach methods are comparable.
6
3. Using the data for our class exercise, test whether Artistic ability vary significantly across sex.
Assume that performance in English is a covariate and the sample is randomly chosen from a
population of normal distribution. Assume all the assumptions for parametric test are met.
a) State the null and alternative hypotheses.
Ho: There is no significant difference in the artistic ability of the students when
grouped according to sex after controlling for the performance in English.
Ha: There is a significant difference in the artistic ability of the students when
grouped according to sex after controlling for the performance in English.
b) Set alpha level.
𝛼 = .05, 2-tailed.
c) Statistical Tool
Analysis of Covariance (ANCOVA)
d) Result
Between-Subjects Factors
Value Label N
1 Male 17
Sex
2 Female 18
Descriptive Statistics
Dependent Variable: Aptitude Score in Artistic Ability
Sex Mean Std. Deviation N
Male 85.47 4.913 17
Female 81.78 3.639 18
Total 83.57 4.635 35
Levene's Test of Equality of Error Variancesa
Dependent Variable: Aptitude Score in Artistic
Ability
F df1 df2 Sig.
5.067 1 33 .031
Tests the null hypothesis that the error variance of
the dependent variable is equal across groups.
a. Design: Intercept + SEX + english + SEX * english
Tests of Between-Subjects Effects
Dependent Variable: Aptitude Score in Artistic Ability
Source Type III Sum of df Mean F Sig. Partial Eta Noncent. Observed
Squares Square Squared Parameter Powerb
Corrected Model 256.940a 3 85.647 5.606 .003 .352 16.817 .913
Intercept 2501.545 1 2501.545 163.730 .000 .841 163.730 1.000
SEX 7.515 1 7.515 .492 .488 .016 .492 .104
english 137.694 1 137.694 9.012 .005 .225 9.012 .828
SEX * english 12.555 1 12.555 .822 .372 .026 .822 .142
Error 473.631 31 15.278
Total 245177.000 35
Corrected Total 730.571 34
a. R Squared = .352 (Adjusted R Squared = .289)
b. Computed using alpha = .05
7
Estimates
Dependent Variable: Aptitude Score in Artistic Ability
Sex Mean Std. Error 95% Confidence Interval
Lower Bound Upper Bound
Male 85.376a .950 83.440 87.313
Female 81.944a .924 80.059 83.828
a. Covariates appearing in the model are evaluated at the following values:
Test Score in English = 83.91.
Pairwise Comparisons
Dependent Variable: Aptitude Score in Artistic Ability
(I) Sex (J) Sex Mean Difference Std. Error Sig.b 95% Confidence Interval for
(I-J) Differenceb
Lower Bound Upper Bound
Male Female 3.454* 1.321 .014 .763 6.144
Female Male -3.454* 1.321 .014 -6.144 -.763
Based on estimated marginal means
*. The mean difference is significant at the .05 level.
Univariate Tests
Dependent Variable: Aptitude Score in Artistic Ability
Sum of df Mean F Sig. Partial Eta Noncent. Observed
Squares Square Squared Parameter Powera
Contrast 103.865 1 103.865 6.836 .014 .176 6.836 .718
Error 486.186 32 15.193
The F tests the effect of Sex. This test is based on the linearly independent pairwise comparisons among the
estimated marginal means.
a. Computed using alpha = .05
c. Adjustment for multiple comparisons: Bonferroni.
e) Decision
Since 𝑝 = .014, and is less than 𝛼 = .05, hence, we reject Ho.
f) Conclusion
The Analysis of Covariance (ANCOVA) reveals that when the students are
grouped according to sex, they differ significantly after controlling for the
performance in English, 𝐹(1, 32) = 6.836; 𝑝 = .014. The result further
indicates that 17.6% of the variance in the artistic ability can be explained by
the sex.
4. An experimental study on the use of manipulative materials was conducted on a group of 15
students using pretest-posttest design. The following table shows the pretest and posttest final
scores of the students. Test the hypothesis that the median score of the students in the pretest is
not significantly different from their median score in the posttest.
Pretest Posttest
72 83
63 76
56 65
80 93
72 85
79 79
70 81
73 86
53 64
83 97
71 85
69 80
65 75
75 87
78 89
a. State the null and alternative hypotheses.
Ho: There is no significant difference in the pretest and posttest median scores of the
students when exposed to the use of manipulative materials.
Ha: There is a significant difference in the
8 pretest and and posttest median scores of the
students when exposed to the use of manipulative materials.
b. Set alpha level.
𝛼 = .05, 2-tailed.
c. Statistical Tool
Wilcoxon Signed Rank Test
d. Result
Descriptive Statistics
Percentiles
Std. Maximu 50th
N Mean Deviation Minimum m 25th (Median) 75th
Pretest 15 70.6000 8.50042 53.00 83.00 65.0000 72.0000 78.0000
Posttest 15 81.6667 9.12349 64.00 97.00 76.0000 83.0000 87.0000
Wilcoxon Signed Ranks Test
Ranks
N Mean Rank Sum of Ranks
Posttest - Pretest Negative Ranks 0a .00 .00
Positive Ranks 14b 7.50 105.00
Ties 1c
Total 15
a. Posttest < Pretest
b. Posttest > Pretest
c. Posttest = Pretest
Test Statisticsa
Posttest - Pretest
Z -3.321b
Asymp. Sig. (2-tailed) .001
a. Wilcoxon Signed Ranks Test
b. Based on negative ranks.
e. Decision
Since 𝑝 = .001, and is less than 𝛼 = .05, hence, we reject Ho.
f. Conclusion
Wilcoxon signed-rank test showed that when the students were exposed to the use
of manipulative, it elicits a significant change on their median scores
(𝑍 = −3.321, 𝑝 = .001). The result indicates that the use of manipulative will
significantly improve the median scores of the students.