# Data Analysis And Findings Education Essay

Published:

This essay has been submitted by a student. This is not an example of the work written by our professional essay writers.

## Chapter 4:

Starting with the research question, we decide which variables in questionnaires can be analyzed to answer the research questions. Based on the number of variables and a type of measured variables (categoricalÂ or continuous ones), it is suitably decided what statistical analysis technique is used. Tools of the analysis are specifically determined in our statistical package (in this case, SPSS). After doing the analysis, with figures and charts, we interpret the output and draw conclusions in relation to the research question.

In proceedings of data analysis, questionnaires are delivered to students and teachers in April and May 2010. Eighty six English students, the undergraduates of ITPC of STU, and one hundred thirty of English teachers, the postgraduates in Tesol of HCMCOU, are working at high schools, colleges, universities nationwide have participated in the survey based on the research objective proposed in the thesis. Class observation conducted in the traditional and virtual class is a necessary way to find out associated factors in a wider context. Data and information are collected and processed. The content is summarized in the following tables.

## 4.1. DATA COLLECTION AND ANALYSIS

## 4.1.1. Class observation

Class observation is a supplementary survey method to help the researcher qualitatively understand context, condition, factors, etc. involved in the research objective.

## 4.1.1.1. Traditional class

The traditional writing courses organized at ITPC of STU has three classes in which each consists of 30-40 graduate students and is instructed by the doctor in English linguistics. First, the instructor presents a theoretical framework of English writing skill. Then, students discuss the topic in pairs and in groups and afterwards, their representatives will show ideas, especially development strategies of the topic. Finally, the lecturer helps students remark the raised ideas, then summarizes the outcome, recommends the notices and gives the assignment at home. The model helps students interact with the instructor and the peers face-to-face. However, delivered knowledge is restricted in the textbook, the syllabus is fixed in the prescribed time, and communicative space is limited in a classroom. In this case, traditional class does not play a main role and is considered as a supplementary constituent to online class.

## 4.1.1.2. Online class

The online writing courses organized at multimedia room of HCMCOU has five classes in which each consists of 30-40 post-graduate students and is instructed by the doctor in computer science. First, the instructor introduces a virtual learning environment and typical distant writing model of previous course that has been completely established here. Then, students are taught to use activities and resources integrated in Moodle and the related issues in utility software. Next, students are assigned in a group of two members, practice by selecting a unit of writing in the coursebook and uploading the document on the internet. At the end of the session, the instructor displays the submission of the groups on the slide screen for students to remark. Finally, the lecturer corrects the students' outcomes, suggests the improvement, and makes careful recommendations of necessary work to them at home. It is easily to recognize that teaching and learning in the online environment helps students reinforce their effort. It enables them to research into the problem independently and creately, and cooperate with partners for a particular purpose to gain knowledge and skill. Therefore, e-learning plays a prominent role in teaching and learning that traditional model did not possibly overcome as knowledge can share among learners worldwide. Online learning can happen anywhere and anytime.

## 4.1.2. Responses to questionnaires

## 4.1.2.1. Questionnaires for students

Questionnaires for students comprise four sections: Information about students (12 questions), Statements about Online Education (12 questions), Statements about Writing Skill (12 questions) and Statements about the website (10 questions). The entity of surveyed students is English students of ITPC of STU that have some understandings in e-learning.

## 4.1.2.1.1. Informations about students

## 4.1.2.1.1.1. Distributed frequency

Table 4.1 shows the informations about of ITPC of STU demonstrated by twelve statements about the gender, age, academic level, etc. Collected data is organized into the table along with frequency and percentage.

In row Position, there are 121 university students (100%). In row Gender, the ratio between female and male have not much difference and 65(53.7%) students are for the former and 35 ones (46.3%) is for the latter. In row Age, there are 102(84.3%) students with the age from 18 to 22. The remainder has the age above 23. Regarding row Years of Post-Secondary Schooling, 27(22.3%) students have been greater than four years to study English and 22(18.2%) for three years, 52(43.0%) for two years and 20(16.5%) for 1 year. Concerning Academic level, there are 121(100%) undergraduate students.

To soundly look into e-learning, it is necessary to consider configuration and network of computer as well as skills of user's computer manipulation. In this part, there are seven questions about a computer, online courses, internet connection, etc.

The number of students who has a computer available at home and at a place of work accounts for very high rate about 118(97.5%) and 80(66.1%) respectively. The number of online courses that is greater than or equal to one is taken with 34 students (45.5%). The approximate number of hours from 1 or more they spend per week for educational purposes and multi-purposes accounts for 112(92.6%) and 61(50.4%) respectively. The number of moderately sufficient computers for the learning needs in the multimedia lab accounts for 78(64.5%). Internet connection and speed for students to study at an average and above average account for 77(63.6%).

## Table 4.1: Informations about students

## Table 4.1: Informations about students (Cont)

## 4.1.2.1.2. Statements about online education

## 4.1.2.1.2.1. Distributed frequency

Contingency table of Moodle's features versus student's opinions for 121 randomly selected students of STU consists of 12 rows and 5 columns in which contain informations on observed frequencies and percentages. Especially, the thirteenth row along with the sixth column is total of rows of the column and total of columns of the row respectively.

Statements about online education include ones about its characters, activities and impact on participants regarding distant training. It contains twelve items that are classified into five groups. They are Modular Layout, Attractiveness, Course Management, Assessment Strategies, Cooperative Learning. In each group, there are many statements. Considering a level from high to low, each question is explored in five levels: "Strongly Agree", "Somewhat Agree", "Neither Agree nor Disagree", "Somewhat Disagree", and "Strongly Disagree".

In the group one, "Modular Layout" in connection with courses' layout comprises template-based models to which content must be added with a level "Somewhat Agree" accounting for 72(59.5%) maximum.

In the group two, "Attractiveness" is involved with "Interface", "Activities", and "Active learning". E-learning's user-friendlly interface with the level "Somewhat Agree" accounts for 69(57.0%). Varied activities and greater time flexibility for students with "Somewhat Agree" accounts for 58(47.9%). Active and dynamic learning with "Somewhat Agree" accounts for 57(47.1%).

In the group three, "Course Management" includes "Education quality", "Managing time", "Re-submitting assignment", "Enrolling a class", "Manipulating activities and resources", and "Manipulating activities and resources" with the level "Somewhat Agree" accounting for 50(41.3%), 57(47.1%), 51(42.1%), 59(48.8%), and 55(45.5%) respectively.

In the group four, "Assessment Strategies" allows e-learning's courses to perform a wide range of response types and peer assessment with the level "Somewhat Agree" accounting for 55(45.5%).

In the group five, "Cooperative Learning" permits students to be divided into subgroups (either visible or separate) and interact with each other synchronously in chat activities, or engage in asynchronous discussions in Wikis and Forums with the level "Somewhat Agree" accounting for 44(36.4%). They can actively collaborate with other students and communicate with instructors during internet activities with the level "Somewhat Agree" accounting for 58(47.9%). (See Table 4.2)

## 4.1.2.1.2.2. Central tendency

In this chapter, at section "Online Education", the student's aptitude as "Strongly Agree", "Somewhat Agree", "Neither Agree nor Disagree", "Somewhat Disagree", and "Strongly Disagree" are encoded in the form of numbers 1, 2, 3, 4, 5 respectively. These values are considered as group midpoints of continuous variable. With such encryption, we enter data in the files *.sav of SPSS, process them and have the following results.

Mean , Median and Mode indicate that typical values tend to lie centrally ("Strongly Agree", "Somewhat Agree") within a set of data arranged according to magnitude.

Standard Deviation is a scale of the scattering of a set of data from its mean and is computed as the square root of variance. For example, with the normal distribution, if the feature "Attractiveness" has mean = 2.2066, one standard deviation =0 .79623, then 68% of students recognize that Moodle's attractiveness is between . That means that their 68% of choice is among the "1" = "Strongly Agree", "2" = "Somewhat Agree", "3" = "Neither Agree nor Disagree".

For resulting values, positive values of skewness indicates that the distribution is skewed to the right, with a longer tail to the right of the distribution maximum. The mass of the distribution is concentrated on the left of the figure (choice). In other words, most students support online education. Figures 4.2a. shows the relative positions of the mean, median, and mode for frequency curves skewed to the right.

For resulting values, positive values of the kurtosis (leptokurtic: "Layout" , "Management", "Coorperation", "Attractiveness", "Assessment") indicate pointed or peaked distributions. Figure 4.3a. shows the kurtosis with k > 0.

## Table 4.3: Statistics on Online Education (Sts)

Modular Layout

Attractiveness

Course Management

Assessment Strategies

Cooperative Learning

N

Valid

121

363

605

121

242

Missing

484

242

0

484

363

Mean

2.2397

2.2066

2.1570

2.1818

2.1942

Std. Error of Mean

.05381

.04179

.03358

.07873

.05268

Median

2.2613a

2.1937a

2.1492a

2.1573a

2.2065a

Mode

2.00

2.00

2.00

2.00

2.00

Std. Deviation

.59196

.79623

.82599

.86603

.81955

Variance

.350

.634

.682

.750

.672

Skewness

-.118

.471

.373

.576

.220

Std. Error of Skewness

.220

.128

.099

.220

.156

Kurtosis

-.445

.461

.082

.597

-.062

Std. Error of Kurtosis

.437

.255

.198

.437

.312

Sum

271.00

801.00

1305.00

264.00

531.00

a. Calculated from grouped data.

## 4.1.2.1.2.3. Histograms

In Chart 4.1, the histogram is a graphical representation of a frequency distribution with adjacent rectangles whose widths show class intervals and whose areas are equal to the corresponding frequencies. The height of a rectangle is identical to the frequency divided by the width of the interval. The total area of the histogram matches with the number of data. The normal curve is also showed on the histogram.

Looking at the graph, we recognize that frequency distributions fall into columns 1, 2, 3. That means that students' opinion focuses on "Strongly Agree", "Somewhat Agree", "Neither Agree nor Disagree". (See Figure 4.4)

## Figure 4.3: The kurtosis with k > 0 and k < 0.

## b) Platykurtic (k < 0)

## a) Leptokurtic (k > 0)

## Figure 4.2: The relative positions of the Mean, Median, and Mode for frequency curves with positive skew and negative skew.

## a)Positively skewed distribution

## b)Negatively skewed distribution

## Chart 4.1: The Histogram of frequence distribution of online education

## 4.1.2.1.2.4. Chi-Square Test online education (Sts)

In the first research question, we recognize the independent variable "Features" in association with the dependent variable "Opinion". To operationalize and hypothesize the variables, we divide the "Features" into five factors: "Modular Layout", "Attractiveness", "Course Management", "Assessment Strategies", "Cooperative Learning" and categorize "Opinion" into five degree: "Strongly Agree", "Somewhat Agree", "Neither Agree nor Disagree", "Somewhat Disagree", "Strongly Disagree". All are arranged in Table 4.1 with rows and columns. The interception between rows and columns is cells where there are observed and expected frequency.

From questionnaires for students, we create contingency table of Moodle's "Features" versus "Opinion" for 86 randomly selected students, then perform chi-square procedure.

## Table 4.4: Chi-Square Tests on online education(Sts)

Value

df

Asymp. Sig. (2-sided)

Pearson Chi-Square

64.785a

44

.022

Likelihood Ratio

72.801

44

.004

Linear-by-Linear Association

.261

1

.609

N of Valid Cases

1452

a. 24 cells (40.0%) have expected count less than 5. The minimum expected count is 1.00.

The following is the procedure to perform a chi-square independent test.

## Assumptions:

24 cells (40.0%) have expected count less than 5.

The minimum expected count is 1.00.

## Step 1 : Hypotheses

Moodle's features and student's opinion are statistically independent.

Moodle's features and student's opinion are statistically dependent.

## Step 2 : Expected frequencies

## ,

where R= row total, C-column total, and n=sample size. See Table 4.4

Step 3 : Check whether the Expected frequencies sastisfy: Assumptions a and b

## Yes

## Yes

## Step 4 : Significance Level

Î± = 0.05

## Step 5 : The critical value

ï£2ï¡,df=(r-1)(c-1)= ï£2 0.05,df=44= 60.48

## Step 6 : Test Statistic and p-value

## Figure 4.5: Criterion for deciding whether or not to reject the null hypothesis

;p-value =0.022 < 0.05

## Step 7 : Conclusion

The value of the Test Statistic is ï£2 = 64.785, which falls in the rejection region (See Figure 4.5 ). Thus, we reject H0.

## Step 8 : State conclusion in words

At the Î± = 0.05 significant level, there is evidence that Moodle's features and student's opinion are statistically dependent.

## 4.1.2.1.2.

## Figure 4.2.5. Correlation and regression of moodle's feature

Using trial-errors and greedy method, we establish table .

## Table 4.5

## Template

## Active_Learning

10.00

20.00

72.00

57.00

39.00

39.00

0.00

2.00

0.00

3.00

50.00

## .

## Table 4.6: Coefficientsa

Model

Unstandardized Coefficients

Standardized Coefficients

t

Sig.

95.0% Confidence Interval for B

Collinearity Statistics

B

Std. Error

Beta

Lower Bound

Upper Bound

Tolerance

VIF

1

(Constant)

4.362

3.895

1.120

.344

-8.035

16.759

Template

.778

.106

.973

7.372

.005

.442

1.114

1.000

1.000

a. Dependent Variable: Active__Learning

## The regression equation

From the above output, the regression equation is: y = 4.362+0.778x

## The coefficient of multiple determination, R2.

## Table 4.7: Model Summaryb

Model

R

R Square

Adjusted R Square

Std. Error of the Estimate

1

.973a

.948

.930

6.57438

a. Predictors: (Constant), Template

b. Dependent Variable: Active_Learning

The coefficient of multiple determination is 0.948; therefore, about 94.8% of the Active_Learning is explained by Template. The regression equation appears to be very useful for making predictions since the value of R2 is close to 1.

## The model is useful for predicting Active_Learning.

## ANOVAb

Model

Sum of Squares

df

Mean Square

F

Sig.

1

Regression

2349.132

1

2349.132

54.350

.005a

Residual

129.668

3

43.223

Total

2478.800

4

a. Predictors: (Constant), Template

b. Dependent Variable: Active_Learning

## Step 1 : Hypotheses

: Î²1 = Î²2 = 0 (Template is not a useful predictor of Active_Learning)

: at least one Î²i â‰ 0 (Template is a useful predictor of Active_Learning)

## Step 2 : Significance Level

Î± = 0.05

## Step 3 : Rejection Region

Reject the null hypothesis if p-value â‰¤ 0.05.

## Step 4 : ANOVA Table (Test Statistic and p-value)

(See Table 4.8) F = 54.350, p-value= 0. 005< 0.05

## Step 5 : Conclusion

Since p-value < 0.05, we shall reject the null hypothesis.

## Step 6 : State conclusion in words

At the Î± = 0.05 significant level, there is evidence that at least one of the predictors (Template) is useful for predicting Active_Learning; Therefore, the model is useful.

## Partial plots

## Chart 4.2: Partial Regression Plots

Active_Learning appears to be linearly related to each of the predictor variables with no visible potential outliers or influential observations (no points away from the main cluster of points); thus, Assumption 1 appears to be satisfied.

## The variables in questions.

We show that the unstandardized residuals (res), studentized residuals (sre), the predicted values (pre), the standard errors of prediction (sep), lower individual confidence interval (lici), upper individual confidence (uici), the lower mean confidence interval (lmci), and upper mean confidence interval (umci) can be found in the data editor window.

## The predictor variables

## Template

## Step1 : Hypotheses

: Î²1 = 0 (Template is not useful for predicting Active_Learning)

: Î²2 â‰ 0 (Template is useful for predicting Active_Learning)

assuming that Template is included in the model

## Step 2 : Significance Level

0.05

## Step 3 : Rejection Region

Reject the null hypothesis if p-value â‰¤ 0.05.

## Step 4 : Test Statistic and p-value

(see Table 4.6) T = 8.481, p-value=0.003 â‰¤ 0.05

## Step 5 : Conclusion

Since p-value â‰¤ 0.05, we shall reject the null hypothesis.

## Step 6 : State conclusion in words

At the Î± = 0.05 significant level, there is evidence that the slope of the Template variable is not zero and, hence, that Template is useful as a predictor of Active_Learning.

## The slopes, , of the population regression line

We are 95% confident that the slope for Template is somewhere between 0.467 and 1.028. In other words, we are 95% confident that for every single-unit increase in Revising, the average Product increases between 0.467 and 1.028.

From output of the data editor window (See Table 4.9), we have

Table 4.9: Output of the data editor window

## Template

## Active Learning

## PRE_1

## RES_1

## SRE_1

## SEP_1

## LMCI_1

## UMCI_1

## LICI_1

## UICI_1

1

10.00

20.00

13.58956

6.41044

1.35123

2.75373

4.82596

22.35316

-5.94377

33.12288

2

72.00

57.00

59.91684

-2.91684

-1.15860

4.87359

44.40690

75.42678

36.56500

83.26868

3

39.00

39.00

35.25877

3.74123

0.79098

2.77814

26.41748

44.10006

15.69046

54.82708

4

0.00

2.00

6.11742

-4.11742

-0.93178

3.25015

-4.22601

16.46084

-14.17388

26.40871

5

0.00

3.00

6.11742

-3.11742

-0.70548

3.25015

-4.22601

16.46084

-14.17388

26.40871

6

50.00

## .

43.47813

3.34431

32.83505

54.12120

23.03246

63.92379

## The point estimate

The point estimate (PRE_1) for the mean Active_Learning, the with 50.00 Template is 43.47813.

## Test the alternative hypothesis

## Step 1 : Hypotheses

: = 43.40 ( when x = 50.00)

: > 43.40 ( when x = 50.00)

## Step 2 : Significance Level

ï¡ = 0.05

## Step 3 : Critical Value(s) and Rejection Region(s)

Critical Value: tÎ±,df = nâˆ’(k+1) = t0.05,df = 3 = t90%CI ,df = 3 =2.353

Reject the null hypothesis if T ï‚³ 2.353 (or if p-value â‰¤ 0.05).

## Step 4 : Test Statistic

= 0.02336

; p-value < 0.05

## Step 5 : Conclusion

Since 0.02336< 2.353 , we shall not reject the null hypothesis.

## Step 6 : State conclusion in words

At the Î± = 0.05 significant level, there is not evidence that the mean Active_Learning with 50.00 Template is greater than 43.40.

## Confidence interval

We are 95% confident that the mean Active_Learning with 50.00 Template is somewhere between 32.83505 (LMCI_1) and 54.12120 (UMCI_1).

## The predicted Active_Learning

The predicted Active_Learning for my class with 50.00 Template is 43.47813.

## Prediction interval

We are 95% certain that the Active_Learning with 50.00 Template will be somewhere between 23.03246 (LICI_1) and 63.92379 (UICI_1).

## 4.1.2.1.3. Statements about online writing skill

## 4.1.2.1.3.1. Distributed Frequency

Statements about Writing Skill in connection with process-based and communicative writing comprise twelve questions that are divided into four stages: Pre-writing, Writing, Editing and Publishing along with related issues. Each stage is divided into more specific statement that is split in five options according to realistic background.

Foundation of educational theory to online writing teaching are as same as traditional one. However, The former has its prominent functions that offer interactive and collaborative activities, a wide range of response and assessment as well as flexibility in submitting an assignment and enrolling in a course. Therefore, statements about online writing are similar to traditional one in the scheme but different in content. (See Table 4.10.)

## 4.1.2.1.3.2. Central tendency and dispersion

With same aforementioned convention and procedure, we obtain observed frequency, percentage of cells in Table 4.11, see the bar chart of online writing in Chart 4.3. , and can easily read the informations on central tendency, dispersion, Distribution in Table 4.11

## 4.1.2.1.3.3. Histograms

Looking at the graph (Chart 4.3), we recognize that frequency distributions fall into columns (1), (2), (3). That means that elements of online writing focus on ones (1), (2), (3).

## Table 4.10: Online Writing Skill

## Table 4.10: Online Writing Skill

## 4.1.2.1.3.4. Chi-square tests on online writing skills (Sts)

In the second research question, we recognize the existence of two independent variables "Online_writing" and "Elements". The variable "Online_writing" is used to show a writing process divided into four stages: Pre-writing, Writing, Editing and Publishing. The variable "Elements" is options used to expressed factors involved in writing process. These alternatives include five choices in which four ones are suggested and the fifth one is reserved for readers to express their ideas. In this part, questionnaires for students are surveyed and a contingency table of online writing and its elements for 121 randomly selected students is set up to perform chi-square procedure. See Table 4.12

## Table 4.11: Statistics on Online Writing (Sts)

Pre_Writing

Writing

Editing

Publishing

N

Valid

544

521

273

232

Missing

0

23

271

312

Mean

2.8015

2.0653

2.4908

2.5259

Std. Error of Mean

.05347

.05310

.06608

.07200

Median

2.8559a

1.8052a

2.5414a

2.4324a

Mode

4.00

1.00

3.00

2.00

Std. Deviation

1.24714

1.21194

1.09179

1.09672

Variance

1.555

1.469

1.192

1.203

Skewness

-.048

.851

-.028

.510

Std. Error of Skewness

.105

.107

.147

.160

Kurtosis

-1.252

-.386

-.916

-.400

Std. Error of Kurtosis

.209

.214

.294

.318

Sum

1524.00

1076.00

680.00

586.00

a. Calculated from grouped data.

## Table 4.12: Chi-Square Tests on online writing skills (Sts)

Value

df

Asymp. Sig. (2-sided)

Pearson Chi-Square

502.465a

44

.000

Likelihood Ratio

463.047

44

.000

Linear-by-Linear Association

18.265

1

.000

N of Valid Cases

1570

a. 1 cells (1.7%) have expected count less than 5. The minimum expected count is 4.92.

At the Î± = 0.05 significant level, there is evidence that online writing and its elements are statistically dependent.

## Chart 4.3: The histogram of frequency distributions of online writing (Sts)

## 4.1.2.1.3.

## Figure 4.2.5. Correlation and regression of online writing skills (Sts)

First, we establish the table of relation of the factors of concerning "Online Writing"