CHAPTER III METHOD AND PROCEDURES 3.1 Data 3.1.

1 The Data Needed The data needed of this study were taken from the students scores that are collected by the test. The source of the data was collected from the Second Grade Students of IPS classes at SMA N-2 Jekan Raya Palangkaraya in Academic year 2006/2007.

3.1.2

Criteria for the Admissibility of the Data The writer took the students answer sheets as the valid out of data. In this study,

the data concern to the students reading comprehension achievements after and without having pre-questioning.

3.1.3

Data Collection and Data Processing Procedures The writer used a test to collect the data. After collecting the data, they were

processed with the following procedures: 1. Editing Find out the data of the Second Grade Students of SMA N-2 Jekan Raya. 2. Coding Each answer sheet is coded by giving code such as A1, A2,B1,B2,and so on in order to keep the students privacy.

23

3. Scoring In this step, the writer gives the score for the students work based on the correct answer that the students make by using the following formula:
SCORE = Number of correc t an swer Number of items tested

X 100 (Depdikbud, 1989:7)

4. Tabulating After scoring the students works, the data will be classified on the table.

3.2 Instrumentation 3.2.1 Instrumentation Development The writer used the test as an instrument to collect the data. There were two kinds of instruments in this study, they were: Instruments for the experimental class and for the control class. The tests were in the form of essay and multiple-choice. The multiplechoice used for the reading comprehension tests and essay used for the Pre-questioning questions. The total numbers of test items for the experimental class were 30 items and 20 test items for the control class. Below are the descriptions of content specification of the test: Table 3.1 Content Specification of Pre-questioning for the experimental class No 1. 2. Content of the tests Pre-questioning before reading for general comprehension Pre-questioning before reading to confirm expectations Total Number of test item 4 items 6 items 10 items Number of item 2, 4, 9, 10 1, 3, 5, 6, 7, 8 10 items

Table 3.2 Content specification of reading comprehension for the Experimental and Control class No 1. 2. Content of test Main Idea Supporting Details Total 3.2.2 Instrumentation Try Out In order to know whether the instrument is suitable or not, the writer tried out the instrument. The try out was held on SMA Negeri-1 Jekan Raya Palangkaraya. The result of try out was useful for the following: 1. To decide the time allocation. 2. To find out whether the instruction is understandable or not. 3. To gather evidence to support the instruments validity and reliability. Before the writer made instrument try out, there were some procedures need to be followed: a. Getting the permissions letter from: The Faculty of Teacher Training and Education of Palangka Raya University Dinas Pendidikan Kota Palangkaraya Number of test item 9 11 20 items Item number 1, 2, 4, 8, 13, 17, 18, 19, 20 3, 5, 6, 7, 9, 10, 11, 12, 14, 15, 16 20 items

b. Asking the permission from the Head Master of SMAN-1 Jekan Raya Palangkaraya and the second grade English teacher in IPS classes. In the process to try out the instruments, the writer found some problems to get the permission letter from the Faculty of Teacher Training and Education of Palangka Raya University. It needed a long time to get the letter. Whereas, the writer needed to conduct the try out as soon as possible. Realizing this reason, the writer had an initiative

to ask the second grade English teacher of IPS classes for the permission to conduct the try out while the letter was in process. Fortunately, the writer got it. Thus, the instrument was tried out to the second year students of IPS classes in SMA Negeri 1 Jekan Raya Palangkaraya. The writer took 2 (two) classes to become the sample of try out. The classes were: XI IPS 2 which consisted of 33 students, and XI IPS 4 which consisted of 35 students. The try out for the instrument without pre-questioning treatment was held on Wednesday, January 17th , 2007 (10.45 - 12.15 am) and with prequestioning treatment was held on Thursday, January 18, 2007 (08.15 - 09.45 am). The time allocation for try out was 60 minutes for each class. In this study, the writer did not measure and use the index of items difficulty, because this study belongs to proficiency test. In other words, the writer only gave the test to measure the students comprehension in text. Based on the result of try out, it can be concluded that: 1. In experimental class, the testees could answer the 30 try out items in 2 x 45 minutes, it means they did one items in 3 minutes. Based on the instruction, the time allocation provided for the sample of the study, was 2 x 30 minutes. So, the time allocation for the experimental class was added to 2 x 45 minutes. In control class, the testees could answer the 20 try out items in 2 x 30 minutes, it means they did one items in 3 minutes. Based on the instruction, the time allocation provided for the sample of the study was 2 x 25 minutes. So, the time allocation for the experimental class was added to 2 x 30 minutes. 2. Based on the result of the try out, the testees could answer the test based on the instruction. As evidence, the students did not ask too much about how to answer the

test. It meant the instruction of the test was not needed to be revised because it was understandable for the students. 3. From the result of the try out, the writer could get the data needed to count the reliability of the instrument. In calculating the result, the writer eliminated two of the students scores in XI IPS 4 from the list to make the equal group. So, the numbers of students from both of class were same.

1.2.3

Instrumentation Validity Arikunto (2002:145), said Sebuah instrumen dikatakan valid apabila mampu

mengukur apa yang diinginkan. Furthermore, Sudjana and Ibrahim (1989:117) have opinion that: Validitas berkenaan dengan ketepatan alat ukur, sehingga betul-betul mengukur apa yang seharusnya diukur. In this study, the writer used two kinds of validity, they are; construct validity and content validity. 3.2.3.1 Construct validity Based on Toendan (2006:133), construct validity refers to the extent to which the results of the data collection process can be interpreted in terms of underlying psychological construct. The instrument has construct validity if there is a relationship between theory and concept of the competence needed to accomplish the tests in this study. In this study, the theories are about pre-questioning and reading comprehension. Those, theories can be seen in Chapter II, Review of related literature, on page 18 (theory of pre-questioning) and on page 6 (theory of reading comprehension). The competence in this study means that the students ability in answering the reading comprehension tests

with pre-questioning treatment and without pre-questioning treatment. The instrument in this study belonged to proficiency test, it meant the test is not measure the students ability after learning process. The form of the reading comprehension tests for experimental and control classes are multiple choices and form pre-questioning for experimental class is essay.

3.2.3.2 Content Validity According to Toendan (2006:132), content validity refers to the extent to which data collection process measures a representative sample of the subject matter or behavior that should be encompassed by the operational definition. Related to this study, the test content is about reading comprehension, about main idea and supporting details, with or without pre-questioning. The following are the tables of content specification of test instrument: Table 3.3 Content Specification of Pre-questioning for the Experimental class No 1. Content of the tests Number of test item Number of item

Pre-questioning before reading 4 items 2, 4, 9, 10 for general comprehension Pre-questioning before reading 2. 6 items 1, 3, 5, 6, 7, 8 to confirm expectations Total 10 items 10 items Table 3.4 Content specification of reading comprehension for the Experimental and Control class No 1. 2. Content of test Main Idea Supporting Details Total 3.2.4 Number of test item 9 11 20 items Item number 1, 2, 4, 8, 13, 17, 18, 19, 20 3, 5, 6, 7, 9, 10, 11, 12, 14, 15, 16, 20 20 items

Instrumentation Reliability

According to Arikunto (2002:154): Reliabilitas menunjuk pada satu pengertian bahwa sesuatu instrumen cukup dapat dipercaya untuk digunakan sebagai alat pengumpul data karena instrumen itu sudah baik. Based on that opinion, the writer measured the reliability of the instrument by using the following formula: rxx = 1VEM Sd 2

Where : rxx = Coefficient Reliability

VEM = Variance Error of Measurement Sd = Standard Deviation

VEM = 1/5 n n = Number of Items (Kasiram, 1984.80)

To find out the standard deviation, the writer used following formula: Sd =
( x x ) n 1
2

Where : Sd = Standard deviation n = Number of students tested = Sum Total

The value will be consulted to the standardization of coefficient reliability. Where: (+0.90) (+1.00) (+0.85) (+0.89) (+0.80) (+0.84) (+0.70) (+0.79) Less than (+0.70) : Excellent Reliability : Very Good Reliability : Good Reliability : Fair Reliability : Poor Reliability ( Soehartono, 1998: 86 ) Based on the result of instrumentation try out, the writer fulfilled them on the table, as follows: Table 3.5 Description of score in instrument try out of experimental class which is used to calculate standard deviation No 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 Students Codes A1 A2 A3 A4 A5 A6 A7 A8 A9 A10 A11 A12 A13 A14 A15 A16 A17 A18 A19 A20 Number of Correct Answer 8 7 9 6 8 5 11 11 14 16 7 9 9 15 11 15 7 9 11 9 Scores (x) 40 35 45 30 40 25 55 55 70 80 35 45 45 75 55 75 35 45 55 35 (x- x ) -10.9 -15.9 -5.9 -20.9 -10.9 -25.9 4.1 4.1 19.1 29.1 -15.9 -5.9 -5.9 24.1 4.1 24.1 -15.9 -5.9 4.1 -15.9 (x- x )2
118.81 252.81 34.81 436.81 118.81 670.81 16.81 16.81 364.81 846.81 252.81 34.81 34.81 580.81 16.81 580.81 252.81 34.81 16.81 252.81

21 22 23 24 25 26 27 28 29 30 31 32 33

A21 A22 A23 A24 A25 A26 A27 A28 A29 A30 A31 A32 A33

13 9 8 7 9 12 14 10 10 12 13 12 12

65 45 40 35 45 60 70 50 50 60 65 60 60 = 1680 x = 50.90

14.1 -5.9 -10.9 -15.9 -5.9 9.1 19.1 -0.9 -0.9 9.1 14.1 9.1 9.1

198.81 34.81 118.81 252.81 34.81 82.81 364.81 0.81 0.81 82.81 198.81 82.81 82.81

= 6472.73

The Number in the table were fulfilled into the standard deviation formula, as follows: Sd =
6472.73 33 1

Sd = Sd =

6472.73 32
202.272

Sd = 14.22

Then, this value was fullfilled into the previous reliability formula, as follow: rxx = 1 VEM Sd 2

1 x 20 rxx = 1 - 5 (14.22) 2

rxx = 1 -

4 202.20

rxx = 1 0,019

rxx = 0,98

The value was consulted to the standardization of coefficient reliability. Where: (+0.90) (+1.00) (+0.85) (+0.89) (+0.80) (+0.84) (+0.70) (+0.79) Less than (+0.70) : Excellent Reliability : Very Good Reliability : Good Reliability : Fair Reliability : Poor Reliability ( Soehartono, 1998: 86 ) Based on the calculation above, the instrumentation try out for the experimental class was categorized into Excellent Reliability.

Table 3.6 Description of score in instrument try out of the control class which is used to calculate standard deviation Number of Correct Answer 12 7 4 4 7 9 10 11 11 10 11 8 7

No 1 2 3 4 5 6 7 8 9 10 11 12 13

Students Codes B1 B2 B3 B4 B5 B6 B7 B8 B9 B10 B11 B12 B13

Scores (x) 60 35 20 20 35 45 50 55 55 50 55 40 35

(x- x ) 16.22 -8.78 -23.78 -23.78 -8.78 1.22 6.22 11.22 11.22 6.22 11.22 -3.78 -8.78

(x- x )2 263.0884 77.0884 565.4884 565.4884 77.0884 1.4884 38.6884 125.8884 125.8884 38.6884 125.8884 14.2884 77.0884

14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33

B14 B15 B16 B17 B18 B19 B20 B21 B22 B23 B24 B25 B26 B27 B28 B29 B30 B31 B32 B33

10 7 9 15 12 10 7 8 11 10 4 7 4 10 9 8 9 9 10 9

50 35 45 75 60 50 35 40 55 50 20 35 20 50 45 40 45 45 50 45 = 1445 x = 43.78

6.22 -8.78 1.22 31.22 16.22 6.22 -8.78 -3.78 11.22 6.22 -23.78 -8.78 -23.78 6.22 1.22 -3.78 1.22 1.22 6.22 1.22

38.6884 77.0884 1.4884 974.6884 263.0884 38.6884 77.0884 14.2884 125.8884 38.6884 565.4884 77.0884 565.4884 38.6884 1.4884 14.2884 1.4884 1.4884 38.6884 1.4884 = 5051.51

The Number in the table were fulfilled into the standard deviation formula, as follows: Sd =
5051.51 33 1

Sd = Sd =

5051.51 32
157.85

Sd = 12.56

Then, this value was fullfilled into the previous reliability formula, as follow: rxx = 1 VEM Sd 2

1 x 20 rxx = 1 - 5 (12.56) 2

rxx = 1 -

4 157.75

rxx = 1 0,025 rxx = 0,97 The value will be consulted to the standardization of coefficient reliability. Where: (+0.90) (+1.00) (+0.85) (+0.89) (+0.80) (+0.84) (+0.70) (+0.79) Less than (+0.70) : Excellent Reliability : Very Good Reliability : Good Reliability : Fair Reliability : Poor Reliability ( Soehartono, 1998: 86 ) Based on calculation above, the instrumentation try out for the control class was categorized into Excellent Reliability.

3.3 3.3.1

Population and Sample Population The population of this study was all the second grade students of SMA Negeri 2

Jekan Raya Palangka Raya in academic year 2006/2007. There were five classes on the second grade in SMA Negeri 2 Jekan Raya Palangka Raya. The following is the description of population of the study:

Table 3.7 The Description of Population No 1. 2. 3. 4. 5. Population Class XI IPA 1 Class XI IPA 2 Class XI IPS 1 Class XI IPS 2 Class Bahasa Total Number of the Population 40 40 38 39 41 198 students

3.3.2

Sample In taking the sample, the writer used the cluster sampling technique. According to

Toendan (2006:93), the cluster sampling is used when it is more feasible or convenient to select groups of individual then it is to select individuals than it is to select individuals from a defined population. Based on the theory of Toendan above, it is clear that the result of this study can be interpreted or generalized to the target population, as long there are some similarities and homogeneities on the sample and target population. But, this technique also has the weaknesses that the results of this study cannot be generalized to the all the SMU in Indonesia, especially in Palangkaraya, and this technique is less accurate than simple random sampling technique. Related to the explanation above, the populations of this study was the sample and it is clear that the sample of this study was the Second Grade Students of IPS classes of SMA N-2 Jekan Raya in Kodya Palangkaraya. The step in taking the sample, first, the writer randomized the three programs: IPA, IPS, and Bahasa. As a result, IPS had been chosen as the sample of the study. From the result of the observation, there were two

classes of IPS that the writer divided into experimental and control classes. The results were XI IPS 1 as the experimental class and XI IPS 2 as the control class.

3.4 Research Methodology 3.4.1 Research Method In this study the writer used the experimental method. According to Toendan (2006: 230) this study belongs to Factorial Experimental Design. The design is as follow: 1 C1 R Where: R C1 C2, 1 2 3 4 X1, X2 O1, O2, O3, O4 3 C2 : Random assignment 4 : Experimental Class : Control Class : Female students of experimental class : Male students of experimental class : Female students of control class : Male students of control class : Treatment using Pre-questioning : Treatment without using Pre-questioning : Stands for an observation or measurement of the subject in the design treatment 2 X1 X1 X2 X2 O1 O2 O3 O4

In this study, the writer investigated: 1. The effect of using pre-questioning and without using pre-questioning on the students reading comprehension achievement. 2. achievement. 3. The effect of treatments with pre-questioning, without pre-questioning and students gender on the reading comprehension. The writer focused his study on the second grade students of IPS classes at SMA N-2 Jekan Raya Palangkaraya in Academic year 2006/2007. The writer gave the treatment to the experimental groups in the form of reading comprehension tests with pre-questioning as the beginning before the students answer the reading comprehension test and the control group had no Pre-questioning as the beginning of the reading comprehension text. The effect of students gender on the students reading comprehension

3.4.2

Variables of the Study There are two variables on this study, that can be explained as follow:

1.

Independent Variable The treatments (with pre-questioning and without pre-questioning) and Students Gender

2.

Dependent Variable Students reading comprehension scores

The relationship between variables:

Treatments Gender

Students reading comprehension achievements

3.4.3

Data Analysis Procedures To analyze the data, the writer followed the procedures:

1. 2.

Checking the students answers and give scores. Analyzing the data by statistic technique of ANOVA using statistical computerized program named SPSS version 13.

3.

Taking the conclusion based on the statistical result. There are some steps should be fulfilled in order to use ANOVA. According to

Usman and Akbar (1995:159): Langkah-langkah menghitung ANOVA: 1. Uji atau asumsikan bahwa data masing-masing dipilih secara acak. 2. Uji atau asumsikan bahwa data masing-masing berdistribusi normal. 3. Uji atau asumsikan bahwa data masing-masing homogen. Moreover, Usman and Akbar (1995:251) said that: sebelum ANOVA digunakan, maka persyaratan yang harus dipenuhi adalah: data setiap kelompok berdistribusi normal, data semua kelompok harus homogen, dan data dipilih secara acak (random). Furthermore, there is a value level whether to know that data were resulted from normal distribution of population and from homogen variance. Safari (2004:84-85), said that: Kaidah penetapan normalitas data: Jika signifikan > 0.50, sampel berasal dari populasi berdistribusi normal.

Jika signifikan < 0.50, sampel berasal dari populasi berdistribusi tidak normal. Kaidah penetapan homogenitas data: Jika signifikan > 0.50, variansi setiap sampel sama (homogen). Jika signifikan < 0.50, variansi setiap sampel tidak sama (tidak homogen). In statistical analysis, the writer used significance based on Saphiro-Wilk formula. Finally, in order to see whether Ho or H1 was accepted, the calculation of the data can be seen in ANOVA table. According to Welkowitz, Ewen, and Cohen (1982:251), the result was consulted to the level significant below: If significant > 0.50, Ho was accepted If significant < 0.50, Ho was rejected

3.5 Validity of the Study There are two validity factors to support this study: 1. External Validity In this study, the writer used cluster-sampling technique, therefore the result of this study can be generalized to the target population, as long as there are some similarities and homogeneity on the sample and target population, but the generalization is less accurate than simple random sampling technique. 2. Internal Validity There are two variables that have strong relationship, cause and effect. Thus, the result of this study can be interpreted. According to Toendan (2006:209) there are several ways to control the treats, as follows: 3.8 Several ways to Control the Treats Threats Description How to control threat

History

Extraneous events occurring at the same time as the treatment may influence performance with regard to the outcome variable It may be the composition of the group rather than the treatment that accounts for performance with regard to the outcome variable

Selection bias

Carefully ascertain that no extraneous events occur Make sure that what happens to one group happens to other(s) Make groups initially equal-preferably by random assignment Pretest to check initial performance of group(s) Randomly assign groups to true experiment Rule out maturation logically (e.g.developmentally implausible) Carefully ascertain that no changes occur in the data collection process If changes do occur, make them happen to both groups Dont select on the basis of extreme scores If you select extreme groups, use random assignment afterwards Strive to prevent dropouts Analyze characteristics of dropouts Use a matching strategy with subsequent random assignment Do not use a pretest Use an unobtrusive pretest Use random assignment to true

Maturation

Improvements between one testing occasion and another may result from routine changer that occur as time passes Differences on various testing occasions may occur because the data collection process has somehow changed

Instrumentation

Statistical regression Groups selected on the basis of extreme score tend to shift toward the mean of the original group when re-tested -

Experimental mortality

Differences on subsequent testing occasions may occur because the composition of the group has changed

Pretesting

The experience of taking the pretest may sensitize subjects to perform better on the posttest

Instability

Socialpsychological threats Expectancy effects

Chance fluctuations in score arising from unreliability may cause changes in performance on various testing occasions Dynamic of the experimental situation may set up alternative treatment that may account for differences Outcomes may occur because experimenter or subjects expected those outcomes

experiment Use reliable tests Use test to significance-after random assignment, if possible Make the experiment as natural and unobtrusive as possible Discourage expectancies Foster opposite rather than supportive expectancies