CHAPTER IV THE EFFECT OF PEER TUTORING STRATEGY ON STUDENTS’ READING COMPREHENSION IN NARRATIVE TEXT AT GRADE X SMA NEGERI 8 PADANGSIDIMPUAN

 

CHAPTER IV

RESEARCH RESULT

To analyze data, researcher has collected data through pre-test and post-test inthe both of classes, experimental class and control class. To find out the effect of Peer Tutoring Strategy on Students’ Reading Comprehension in Narrative Text, researcher has calculated data by using quantitative analysis. Researcher used the formulation of t-test to test hypothesis. Next, researcher described data as follow:

A.    Description of Data

1.      Description Data of Pre-test

a.      Pre-test Score of Experimental Class

Researcher took class X MIA 2 as the experimental class. Based on students’ answer in pre-test, researcher has calculated the students’ score in appendix 13 and 14. Then, researcher drawn the table sum in the following:

Table 5

Score of Experimental Class in Pre-test

Total	1840
Highest score	85
Lowest score	50
Mean	70.74
Median	81.3
Modus	61.5
Range	35
Interval	6
Standard deviation	27.66
Variant	-74.61

Based on the table above the total score of experimental class in pre-test was 1840, mean was 70.74, median was 81.3, modus was 61.5, range was 35, interval was 6, standard deviation was 27.66, and variant was -74.61.The researcher got the highest score was 85 and the lowest score was 50. Then, the calculation of the frequency distribution of the students’ score in experimental class can be applied into table frequency distribution as follow:

Table 6

Frequency Distribution of Experimental Class in Pre-test

No	Interval	Mid Point	F	Percentages
1	50-55	52.5	3	11.5%
2	56-61	58.5	3	11.5%
3	62-67	64.5	3	11.5%
4	68-73	70.5	4	15.4%
5	74-79	76.5	7	26.9%
6	80-85	82.5	6	23.1%
i=6		-	26	100%

 

From the table above, it can be concluded that the most students are in interval  74 – 79 (7 students/26.9%). The least of students is 50 – 55(3 student/11.5%). Clear description of the data is presented in histogram on the following figure:

 

 

 

 

 

 

 

 

 

 

 

Figure 1. Histogram Result Score of Students' Reading Comprehension in Narrative Text In Experimental Class (Pre-Test)

Based on the figure above, the frequency of students’ score from 50 up to 55 was 3; 56 up to 61 was 3; 62 up to 67 was 3; 68 up to 73 was 4; 74 up to 79 was 7; 80 up to 85was 6. The histogram shows that the highest interval (74 - 79) was 7 students, and the lowest interval (50-55) was 3 students.

b.      Pre-Test Score of Control Class

In pre-test of control class, researcher calculated the result that had been gotten by the students in answering the test. Score of pre-test control class can be seen in the following table:

 

 

 

Table 7

 Score of Control Class in Pre-Test

Total	1855
Highest score	85
Lowest score	55
Mean	73.34
Median	82
Modus	69.5
Range	30
Interval	5
Standard deviation	18.5
Variant	97.12

Based on the table above the total score of control class in pre-test was 1855, mean was 73.34, standard deviation was 18.5, variant was 97.12, range was 30, interval was 5, median was 82 and modus was 69.5. Researcher got the highest score was 85 and the lowest score was 55. It can be seen on appendix 13 and 14. Then, the computed of the frequency distribution of the students’ score of control class can be applied into table frequency distribution as follow:

Table 8

Frequency Distribution of Control Class in Pre-Test

No	Interval	Mid Point	Frequency	Percentages
1	55-59	57	3	11.5%
2	60-64	62	3	11.5%
3	65-69	67	3	11.5%
4	70-74	72	4	15.3%
5	75-79	77	5	19.2%
6	80-84	82	4	15.3%
7	85-89	87	4	15.3%
i = 5		-	26	100%

 

From the table above, it can be concluded that the interval (75 – 79) had the biggest frequency (5 students/19.2%). The highest interval (85– 89) had 4 students and the lowest interval was(55 – 59) with 3 students.In order to get description of the data clearly and completely, the researcher presents them in histogram on the following figure:

 

 

 

 

Figure 2.Histogram Result Score of Students' reading comprehension in narrative text in Control Class (Pre-test)

 

 

Based on the figure above, the frequency of students’ score from 55 up to 59 was 3; 60 up to 64 was 3; 65 up to 69 was 3; 70 up to 74 was 4; 75 up to 79was 5; 80 up to 84 was 4; and 85 up to 89 was 4.

 

 

 

2.      Description Data of Post-test

a.      Post-Test Score of Experimental Class

The calculation of the result that had been gotten by the students in answering the question (test) after researcher did the treatment by using Peer Tutoring can be seen in the following table:

Table 9

Score of Experimental Class in Post Test

Total	2135
Highest score	100
Lowest score	60
Mean	82.12
Median	91
Modus	71.17
Range	40
Interval	7
Standard deviation	23.87
Variant	101.34

Based on the above table the total score of experiment class in post-test was2135, mean was 82.12, standard deviation was 23.87, variant was 101.34, median was 91, range was 40, modus was 71.17, and interval was 7. The students’ highest score was 100 and the lowest score was 60. It can be seen on appendix 16 and 17. Then, the calculation of the frequency distribution of the students’ score of experiment class can be applied into table frequency distribution as follow:

 

 

 

Table 10

Frequency Distribution of Experimental Class in Post-test

No	Interval	Mid Point	Frequency	Percentages
1	60-66	63	3	11.5%
2	67-73	70	3	11.5%
3	74-80	77	4	15.3%
4	81-87	84	8	30.7%
5	88-94	91	4	15.3%
6	95-101	98	4	15.3%
i= 7		-	40	100%

 

From the table above, it can be concluded that the middle interval (81–87) had the biggest frequency (8students/30.7%). The highest interval (95-101) had 4 students and the lowest interval was(60 – 66) with 3 students.Researcher presented them in histogram as follow:

Figure 3. Histogram Result Score of Students' Reading Comprehension in Narrative text in Experimental Class (Post-test)

           

 

Based on the figure above, the frequency of students’ score from 60up to 66 was 3; 67 up to 73 was  3; 74 up to 80 was 4; 81 up to 87 was 8; 88 up to 94 was 4; 95 up to 101 was 4. Then, the interval which had highest frequency was 81 - 87 and the interval which had lowest frequency was 60 – 66.

b.      Post-Test Score of Control Class

As the control class, researcher took class X MIA 1. The result that had been gotten by the students in answering the question (test) after researcher taught the reading by using lecturer strategy can be seen in the following table:

Table 11

The Score of Control Class in Post-Test

Total	1980
Highest score	95
Lowest score	55
Mean	75.76
Median	77.46
Modus	72
Range	40
Interval	7
Standard deviation	22.54
Variant	120.61

    Based on the above table the total score of control class in post-test was 1980, mean was 75.76, standard deviation was 22.54, variant was 120.61, median was 77.46, range was 40, modus was 72, and interval was 7.  Researcher got the highest score was 95 and the lowest score was 55. It can be seen on appendix 16 and 17. Then, the computed of the frequency distribution of the students’ score of control class can be applied into table frequency distribution as follow:

 

Table 12

Frequency Distribution of Control Class in Post-test

No	Interval	Mid Point	Frequency	Percentages
1	55 – 61	58	3	11.5%
2	62 – 68	65	4	15.4%
3	69 – 75	72	7	26.9%
4	76 – 82	79	4	15.4%
5	83 – 89	86	4	15.4%
6	90 – 96	93	4	15.4%
i = 7		-	26	100%

 

From the table above, it can be concluded that the middle interval (69–75) had the biggest frequency (7students/26.9%). The highest interval (90-96) had 4 students and the lowest interval was(55 – 61) with 3 students.For the clear description of the data, researcher presents them in histogram on the following figure:

Figure 4. Histogram Result Score of Students' Reading Comprehension in Narrative text in Control Class (Post-test)

                                   

 

Based on the figure above, the frequency of students’ score from 55 up to  61 was 3; 62 up to 68 was  4; 69 up to 75 was 7; 76 up to 82 was 4; 83 up to 89 was 4; 90 up to 96 was 4. Then, the interval which had highest frequency was 69 - 75 (7 students) and the interval which had lowest frequency was 55 – 61 (3 students).

3.    Description of Data Comparison between Pre-Test and Post-Test ofExperimental and Control Class

a.      Comparison Data between Pre-test and Post-test Experimental Class

Based on the description data in pre-test of experimental and control class, there was comparison score between pre-test experimental class before and after gave a treatment by peer tutoring strategy . It can be seen in the following table:

Table 13

Comparison Score of Students' Reading Comprehension

in Narrative Text

in Pre-test and Post-test (Experimental Class)

No	Interval	Pre-Test		Post-Test
		Frequency	Score	Frequency	Score
1	50-55	3	50,55,55	0	-
2	56-61	3	60,60,60	2	60,60
3	62-67	3	65,65,65	1	65
4	68-73	4	70,70,70,70	3	70,70,70
5	74-79	7	75,75,75,75,75,75,75	3	75,75,75
6	80-85	6	80,80,85,85,85,85	9	80,85,85, 85,85,85, 85,85
7	86-91	0	-	4	90,90,90, 90
8	92-97	0	-	1	95
9	98-103	0	-	3	100,100, 100

From the table above, it can be seen that the interval 80 – 89 had the highest frequency in pre-test, it was 9 students. While in post-test the highest frequency was on interval 70 – 79, it was 6 students. The lowest frequency in pre-test was on interval 90 – 99 and 100-109, it was 0. While the lowest frequency in post-test was on interval 40-49, it was 1.

In order to get description of the data clearly and completely, the researcher presents them in histogram on the following figure:

Figure 5.Histogram Comparison Data of Students' Reading Comprehension in Narrative text in Pre-test and Post-test (Experimental Class)

 

 

Based on the figure above, the frequency of students’ score in pre test of experimental class from 50 up to 55 was 3; 56 up to 61 was 3; 62up to 67was 3; 68 up to 73 was 4; 74 up to 79 was 7; 80 up to 85 was 6; 86 up to 91 was 0, 92 up to 97 was 0, and 98 up to 103 was 0. Meanwhile, the frequency of students’ score in post test from 50 up to 55 was 0; 56 up to 61 was 2; 62 up to 67 was 1; 68 up to 73 was 3; 74 up to 79 was 3; 80 up to 85 was 9; 86 up to 91 was 4, 92 up to 97 was 1, and 98 up to 103 was 3. Then, the interval which had highest frequency in pre test was 74-79(7 students) and the interval which had lowest frequency was 98-103 (0student). In post test of experimental class, the interval which had highest frequency was 80 - 85 (9 students) and the interval which had lowest frequency was 50-55 (0 student).

b.      Comparison Data between Pre-test and Post-test Control Class

Based on the description data in pre-test and post-test of control class, there was the comparison score between pre-test control class before and after gave a treatment by using lecture strategy. It can be seen in the following table:

Table 14

Comparison Score of Students’ Reading Comprehension

in Narrative Text

in Pre-test and Post-test (Control Class)

No	Interval	Pre-Test		Post-Test
		Frequency	Score	Frequency	Score
1	55 – 61	6	55,55,55, 60,60,60	3	55,60,60
2	62 – 68	3	65,65,65	4	65,65,65,65
3	69 – 75	9	70,70,70, 70,75,75, 75,75,75	7	70,70,75,75, 75,75,75
4	76 – 82	4	80,80,80, 80	4	80,80,80, 80
5	83 – 89	4	85,85,85, 85	4	85,85,85,85
6	90 – 96	0		4	90,90,95,95

From the table above, it can be it can be seen that the interval 69 –75 had the highest frequency in pre-test (9 students) and that interval 69 –75had the highest frequency in post-test (7 students). The lowest frequency in pre-test was on interval 62 – 68, it was 3 students. While the lowest frequency in post-test was on interval 55 – 61, it was 3 students.

For the clear description of the data, the researcher presents them in histogram on the following figure:

Figure 6.Histogram Comparison Data of Students' Reading Comprehension in Narrative Text in Pre-test and Post-test (Control Class)

 

Based on the figure above, the frequency of students’ score of

Control class in pre-test from 55up to 61was 6; 62 up to 68 was 3; 69 up to 75 was 9; 76 up to 82 was 4; 83 up to 89 was 4; 90 up to 96 was 0. Meanwhile, the frequency of students’ score of control class in post-test from 55 up to 61 was 3; 62 up to 68 was 4; 69 up to 75 was 7; 76up to 82 was 4; 83 up to 89 was 4; 90 up to 96 was 4. Then, the interval which had highest frequency in pre test was 69-75 (9 students) and the interval which had lowest frequency was 90– 96 (0 student). In post test of control class, the interval which had highest frequency was 69 - 75 (7 students) and the interval which had lowest frequency was 55-61  (3 students).

c.       Comparison Data between Pre-test Experimental and Control Class

   By giving pre test to both of classes the researcher knew the students’ reading comprehension before givingthe treatment.In pre test, the researcher did not apply treatment to experimental and control class. The comparison data of pre-test between experimental and control class can be seen on the following table:

Table 15

Comparison Score of Students’ Reading Comprehension

in Narrative Text

in Experimental and Control Class (Pre-test)

No.	Interval	Frequency
		Control	Experimental
1	50-55	3	3
2	56-61	3	3
3	62-67	3	3
4	68-73	4	4
5	74-79	5	7
6	80-85	8	6

From the table above, it can be seen that the interval 80 – 85 (8 students) the highest frequency in control and the interval 74-79 (7 students) the highest frequency in experimental class. The lowest frequency in control class that interval 50-55 (3 students) and lowest frequency in experimental class was on interval 50-55 (3 students).

For the clear description of the data is presented in the histogram comparison data pre-test between experimental and control class on the following figure:

Figure 7.Histogram Comparison between Description Data of Students' Reading Comprehension in Narrative Text in Experimental and Control Class (Pre-Test)

 

Based on the figure above, the frequency of students’ score of control class in pre-test from 50 up to 55 was 3; 56 up to 61 was  3; 62 up to 67 was 3; 68 up to 73 was 4; 74 up to 79 was 5; 80 up to 85 was 8. Meanwhile, the frequency of students’ score of experimental class in pre-test from 50 up to 55 was 3; 56 up to 61 was  3; 62 up to 67 was 3; 68 up to 73 was 4; 74 up to 79 was 7; 80 up to 85 was 6.Then, the interval 80 – 85 (8 students) the highest frequency in control and the interval 74-79 (7 students) the highest frequency in experimental class. The lowest frequency in control class that interval 50-55 (3 students) and lowest frequency in experimental class was on interval 50-55 (3 students).

d.      Comparison Data between Post-test Experimental and Control Class

 

The researcher got the comparison data between post-testscore in experimental and control class after giving the treatment. The comparison datacan be seen on the following table:

Table 16

Comparison Score of Students’ Reading Comprehension

in Narrative Text

in Experimental and Control Class (Post-test)

No.	Interval	Frequency
		Control	Experimental
1	55 – 61	3	2
2	62 – 68	4	1
3	69 – 75	7	6
4	76 – 82	4	1
5	83 – 89	4	8
6	90 – 96	4	5
7	97-103	0	3

From the table above, it can be seen that the interval 69 – 75 has the highest frequency in pre-test, it was 7 students in control class. While the highest frequency in experimental class was on interval 83 – 89, it was 8 students. The lowest frequency in control class was on interval 97 –103, it was 0. While the lowest frequency in experimental class was on interval 62 – 68, it was 1.

For the clear description of the data is presented in the histogram comparison between description data post test of experimental andcontrol class on the following figure:

Figure 8.Histogram Comparison between Description Data of Students' Reading Comprehension in Narrative Text in Experimental and Control Class (Post-test)

           

 

Based on the figure above, the frequency of students’ score of control class in post test from 55 up to 61 was 3; 62 up to 68 was  4; 69 up to 75 was 7; 76 up to 82 was 4; 83 up to 89 was 4; 90 up to 96 was 4; 97 up to 103 was 0. Meanwhile, the frequency of students’ score of experimental class in post test from 55 up to 61 was 2; 62 up to 68 was  1; 69 up to 75 was 6; 76 up to 82 was 1; 83 up to 89 was 8; 90 up to 96 was 5; 97 up to 103 was 3. Then, the interval 69 – 75has the highest frequency in pre-test, it was 7 students in control class. While the highest frequency in experimental class was on interval 83 – 89, it was 8 students. The lowest frequency in control class was on interval 97-103, it was 0. While the lowest frequency in experimental class was on interval 62 – 68, it was 1.

From the description of comparison data above, it can be conluded that the students’ scores of experimental class by using peer tutoring was higher than the students’ score of control class by using lecture strategy.

 

B.  Technique of Data Analysis

1.    Requirement Test

a.    Normality and Homogeneity of Experimental and Control Class in Pre-Test

Table 17

Normality and Homogeneity in Pre-Test

Class	Normality Test		Homogeneity Test
	x2count	x2table	fcount	ftable
Experimental Class	6.07	11.070	-1.30<4.28
Control Class	2.53	11.070

 

Based on the table above, the score of experiment class Lo = 6.07< Lt = 11.070 with n = 26 and control class Lo = 2.53< Lt = 11.070 with n= 26, and real level 0.05. Cause Lo< Lt in the both class. So, H_a was accepted. It means that experiment class and control class were distributed normal. It can be seen in appendix 14.

The coefficient of F_count= -1.30was compared with F_table. Where F_table was determined at real α 0.05, and the different numerator dk = N-1 = 26-1=25 and denominator dk N-1 = 26–1 = 25. So, by using the list of critical value at F distribution is got F_0.05 = 4.28. It showed that F_count-1.30<F_table4.28. It showed that both experimental and control class were homogeneous.The calculation can be seen on the appendix 15.

b.   Normality and Homogeneity of Experimental and Control Class in Post-Test

Table 18

Normality and Homogeneity in Post-Test

Class	Normality Test		Homogeneity Test
	x2count	x2table	fcount	ftable
Experimental Class	1.83	11.070	0.84<4.28
Control Class	0.34	11.070

 

The previous table shows that the score of experimental class Lo = 1.83< Lt = 11.070 with n = 26 and control class Lo = 0.34< Lt = 11.070 with n = 26, and real level 0.05. Because Lo< Lt in the both class, it means H_a was accepted. It meants that experiment class and control class were distributed normal. The calculation can be seen in appendix 17.

The coefficient of F_count= 0.84was compared with F_table. Where F_table was determined at real α 0.05, and the different numerator dk = N-1 = 26-1 = 25 and denominator dk N-1 = 26-1 = 25. So, by using the list of critical value at F distribution is got F_0.05 =4.28. It showed that F_count0.84<F_table4.28. So, the researcher concluded that the variant from the data of the students’ reading comprehension in narrative text at grade Xof SMANegeri8 Padangsidimpuan in experimental and control class was homogenous. The calculation can be seen on the appendix 18.

2.    Hypothesis Test

After calculating the data of post-test, researcher has found that post-test result of experimental and control class is normal and homogenous. The data would be analyzed to prove the hypothesis. It used formula of t-test. Hypothesis of the research was “peer tutoring has significant effect on students’ reading comprehension in narrative text at gradeX ofSMANegeri8 Padangsidimpuan”. The calculation can be seen on the appendix 19 and 20. The result of t-test was as follow:

Table 19

Result of T-test from the Both Averages

Pre-test		Post-test
tcount	ttable	tcount	ttable
-2.92	2.021	2.28	2.021

 

The test hypothesis have two criteria. First, if t_count<t_table, H₀ is accepted. Second, t_count>t_table, H_a is accepted. Based on researcher calculation in pre test, researcher found that t_count -2.92 while t_table 2.021with opportunity (1 –α ) = 1 - 5% = 95% and dk = n₁ + n₂ – 2 = 26 + 26 – 2 = 50. Cause t_count<t_table(-0.29<2.021), it means that hypothesis H_a was rejected and H₀ was accepted. So, in pre test, the two classes were same. There is no difference in the both classes. But, in post test, researcher found that t_count2.28while t_table 2.021 with opportunity (1 –α ) = 1 - 5% = 95% and dk = n₁ + n₂ – 2 = 26 + 26 – 2 = 50. Cause t_count>t_table (2.28> 2.021), it means that hypothesis H_a was accepted and H₀ was rejected. The calculation can be seen on the appendix 20. In this case, the mean score of experimental class by using peer tutoringstrategywas 82.12 and mean score of control class was 75.76 that was taught by using lecturer strategy. So, there was the significant effect of using peer tutoring strategy on students’ reading comprehension in narrative text at grade XSMA Negeri 8 Padangsidimpuan.

 

C.    Discussion

Researcher discussed result of this research and compared with the theory and the related findings that have been stated by the researcher. First, Related to the theory from Densereau stated that Peer tutoring has been used with college students to develop higher-order skills such as reading comprehension.[1] L. S. Fuchs Fuchs also said the peer tutoring strategy has been well validated for promoting the development of low-level skills, such as spelling, math and reading.[2] Then, Anita Lie said that teaching by peers (peer tutoring) was more effective than teaching by teachers, this is due to the background of the student experience that is similar to one another than with the schemata teachers. It means that learning process does not have come from the teacher to students, but can also from students to other students.

Second, based on related findings: Yanuar Irakas Prihatno has done research.[3] He found that students’ scores mean of the pretest was 66.26, whose the highest and lowest score in turn were 88, and 32, in the post-test mean score was 78.5 having 85 as the biggest, and 55 as the smallest score. It indicated that the students’ scores improved during the implementation of peer tutoring.

Then, Dika Prima Nugraha Putra also has done research.[4] The result of the research had shown that Peer Tutoring could be an effective technique to improve the students’ reading comprehension. This conclusion came from the analysis of the test result. He found that students’ scores mean of the pre-test was 47.26, whose the highest and lowest score in turn were 70, and 35, in the post-test having mean score was 77.41 and having 85 as the biggest, and 55 as the smallest score.

The last, Besty Maghfiroh has also done research.[5] The result of the research had shown that Peer Tutoring could be an effective method to increase the students’ interest and matematika study result. This conclusion came from the analysis of the test result. He found that students’ scores mean of the pre-test was 61.83, in the post-test having mean score was 77.30.

The research result and the theory have proven that this strategy is good where the students were so enthusiastic to follow the lesson. This proofs show that peer tutoringis suitable to be applied in teaching reading comprehension. So, peer tutoring strategy has given the effect to the research that has been done by the researcher or the other researcher who mentioned in related findings.

 

D.    Limitation of the Research

The research was limited in some situations. It was the problems in the class that appeared during doing the research, but the researcher couldn’t hold or improve those things. The limitation of the research was as follow:

1.      The researcher was not sure whether all of students in the experimental class and control class did the test honestly. There was a possibility that some of them answered the test by copying or imitating their friends’ answer.

2.      The students were noisy while in learning process. They were not concentrating in following the learning process. Some of them talked to their friends and some of them did something outside the teacher’s rule. Of course it made them cannot get the teacher’s explanation well and gave the impact to the post-test answer.

3.      It was also a possibility that some of students were not too serious in answering the pre-test and post-test. It may caused by the test, because they knew before that the test would not influence their score in the school. It made them answer the test without thinking hard and the answer of the test was not pure because they did not do it seriously.

 

---

[1]D. F. Dansereau, Transfer from cooperative to individual studying. Journal of Reading, (5),1987, p. 614-619, assessed from goo.gl/uZUilo, retrieved on  December 2017 at 08.30 a.m.

.

[2]L. S. Fuchs Fuchs, D., Phillips, N. B., Hamlett, C. L., &Karns, K. Acquisition and transfer effects of classwide peer-assisted learning strategies in mathematics for students with varying learning histories School Psychology Review, (24), 1995, p. 604-620, assessed from https://eric.ed.gov/?id=EJ

589164, retrieved on  December 2017 at 09.06 a.m.

[3]Yanuar Irakas Prihatno, Improving Grade X Students’ Reading Comprehension by Using Peer Tutoring at MAN III Yogyakarta”, 2014, accessed from http://eprints.uny.ac.id/19826/1/ Yanuar%20 Irakas%20Prihatno%2009202244030.pdf.  Retrieved on  2017 at 08.15 a.m.

[4]Dika Prima Nugraha Putra, Improving Grade XI Students’ Reading Comprehension by Using Peer Tutoring at SMA N 1 Pleret, 2013, accessed from http//:eprints.umk.ac.id/1173/1/ TITLE.pdf. Retrieved on  2017 at 08.20 a.m.

[5]Besty Maghfiroh, Pengaruh Strategi Pembelajaran Kooperatif Model Tutor Sebaya (Peer Tutoring) Terhadap Minat dan Hasil Belajar Matematika Kelas V Madrasah Ibtidaiyah Darul Huda Ngaglik Bandulan Sleman Yogyakarta, 2012, accessed from http://digilib.uin-suka.ac.id/9948/ 1/BAB%20I%2C%20 IV%2C%20DAFTAR%20PUSTAKA.pdf. Retrieved on  2017 at 08.25 a.m.