Docstoc

Data Collection and Data Processing Procedures

Document Sample
Data Collection and Data Processing Procedures Powered By Docstoc
					                                     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:

                 Number of correc t an swer
    SCORE =                                 X 100
                  Number of items tested

                                                                        (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 multiple-

choice 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

                                                Number of test
    No            Content of the tests                               Number of item
                                                   item
             Pre-questioning before reading
     1.                                             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 3.2 Content specification of reading comprehension for the Experimental and
            Control class

No         Content of test      Number of test item               Item number
1.           Main Idea                 9                  1, 2, 4, 8, 13, 17, 18, 19, 20
                                                          3, 5, 6, 7, 9, 10, 11, 12, 14,
 2.       Supporting Details              11
                                                                      15, 16
                Total                 20 items                       20 items


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

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 pre-

questioning 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.



    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

                                                 Number of test
      No          Content of the tests                                Number of item
                                                    item
         Pre-questioning before reading
      1.                                      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       Content of test        Number of test item             Item number
1.      Main Idea                      9                1, 2, 4, 8, 13, 17, 18, 19, 20
                                                        3, 5, 6, 7, 9, 10, 11, 12, 14,
 2.     Supporting Details               11
                                                                  15, 16, 20
                Total                 20 items                     20 items


3.2.4      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:

           VEM
rxx = 1-
            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:

                        2
           Σ( x − x )
Sd =
              n −1



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)         : Excellent Reliability

(+0.85) – (+0.89)         : Very Good Reliability

(+0.80) – (+0.84)         : Good Reliability

(+0.70) – (+0.79)         : Fair Reliability

Less than (+0.70)         : 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

                                Number
              Student’s
    No
               Codes
                               of Correct      Scores (x)    (x- x )             (x- x )2
                                Answer
     1           A1                 8               40        -10.9              118.81 
     2           A2                 7               35        -15.9              252.81 
     3           A3                 9               45         -5.9              34.81 
     4           A4                 6               30        -20.9              436.81 
     5           A5                 8               40        -10.9              118.81 
     6           A6                 5               25        -25.9              670.81 
     7           A7                11               55         4.1               16.81 
     8           A8                11               55         4.1               16.81 
     9           A9                14               70        19.1               364.81 
    10           A10               16               80        29.1               846.81 
    11           A11                7               35        -15.9              252.81 
    12           A12                9               45         -5.9              34.81 
    13           A13                9               45         -5.9              34.81 
    14           A14               15               75        24.1               580.81 
    15           A15               11               55         4.1               16.81 
    16           A16               15               75        24.1               580.81 
    17           A17                7               35        -15.9              252.81 
    18           A18                9               45         -5.9              34.81 
    19           A19               11               55         4.1               16.81 
    20           A20                9               35        -15.9              252.81 
    21               A21          13           65             14.1            198.81 
    22               A22           9           45              -5.9           34.81 
    23               A23           8           40             -10.9           118.81 
    24               A24           7           35             -15.9           252.81 
    25               A25           9           45              -5.9           34.81 
    26               A26          12           60              9.1            82.81 
    27               A27          14           70             19.1            364.81 
    28               A28          10           50              -0.9            0.81 
    29               A29          10           50              -0.9            0.81 
    30               A30          12           60              9.1            82.81 
    31               A31          13           65             14.1            198.81 
    32               A32          12           60              9.1            82.81 
    33               A33          12           60              9.1            82.81 
                                            Σ = 1680                       ∑ = 6472.73
                                            x = 50.90


The Number in the table were fulfilled into the standard deviation formula, as follows:

                 6472.73
         Sd =
                  33 − 1

                 6472.73
         Sd =
                   32

         Sd =    202.272

         Sd = 14.22



Then, this value was fullfilled into the previous reliability formula, as follow:

                     VEM
         rxx = 1 -
                      Sd 2

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

                       4
         rxx = 1 -
                     202.20
         rxx = 1 – 0,019

         rxx = 0,98



The value was consulted to the standardization of coefficient reliability.

Where:

(+0.90) – (+1.00)          : Excellent Reliability

(+0.85) – (+0.89)          : Very Good Reliability

(+0.80) – (+0.84)          : Good Reliability

(+0.70) – (+0.79)          : Fair Reliability

Less than (+0.70)          : 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
               Student’s
    No
                Codes
                                of Correct      Scores (x)   (x- x )            (x- x )2
                                 Answer
     1            B1                12               60      16.22                263.0884
     2            B2                 7               35       -8.78                77.0884
     3            B3                 4               20      -23.78               565.4884
     4            B4                 4               20      -23.78               565.4884
     5            B5                 7               35       -8.78                77.0884
     6            B6                 9               45       1.22                  1.4884
     7            B7                10               50       6.22                 38.6884
     8            B8                11               55      11.22                125.8884
     9            B9                11               55      11.22                125.8884
    10            B10               10               50       6.22                 38.6884
    11            B11               11               55      11.22                125.8884
    12               B12          8            40             -3.78             14.2884
    13               B13          7            35             -8.78             77.0884
    14               B14          10           50             6.22              38.6884
    15               B15          7            35             -8.78             77.0884
    16               B16          9            45             1.22               1.4884
    17               B17          15           75            31.22             974.6884
    18               B18          12           60            16.22             263.0884
    19               B19          10           50             6.22              38.6884
    20               B20          7            35             -8.78             77.0884
    21               B21          8            40             -3.78             14.2884
    22               B22          11           55            11.22             125.8884
    23               B23          10           50             6.22              38.6884
    24               B24          4            20            -23.78            565.4884
    25               B25          7            35             -8.78             77.0884
    26               B26          4            20            -23.78            565.4884
    27               B27          10           50             6.22              38.6884
    28               B28          9            45             1.22               1.4884
    29               B29          8            40             -3.78             14.2884
    30               B30          9            45             1.22               1.4884
    31               B31          9            45             1.22               1.4884
    32               B32          10           50             6.22              38.6884
    33               B33          9            45             1.22               1.4884
                                            Σ = 1445                       ∑ = 5051.51
                                            x = 43.78


The Number in the table were fulfilled into the standard deviation formula, as follows:

                 5051.51
         Sd =
                  33 − 1

                 5051.51
         Sd =
                   32

         Sd = 157.85

         Sd = 12.56



Then, this value was fullfilled into the previous reliability formula, as follow:

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

                       4
         rxx = 1 -
                     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)             : Excellent Reliability

(+0.85) – (+0.89)             : Very Good Reliability

(+0.80) – (+0.84)             : Good Reliability

(+0.70) – (+0.79)             : Fair Reliability

Less than (+0.70)             : Poor Reliability

                                                                    ( Soehartono, 1998: 86 )

         Based on calculation above, the instrumentation try out for the control class was

categorized into “Excellent Reliability”.



3.3     Population and Sample

3.3.1    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              Population                            Number of the
                                                        Population
 1.     Class XI IPA 1                                     40
 2.     Class XI IPA 2                                     40
 3.     Class XI IPS 1                                     38
 4.     Class XI IPS 2                                     39
 5.     Class Bahasa                                       41
                    Total                                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               X1       O1

                              C1
                                               2               X1       O2

        R
                                               3               X2       O3
Where:
                              C2
R                         : Random assignment
                                            4                  X2       O4
C1                        : Experimental Class

C2,                       : Control Class

1                         : Female students of experimental class

2                         : Male students of experimental class

3                         : Female students of control class

4                         : Male students of control class

X1,                       : Treatment using Pre-questioning

X2                        : Treatment without using Pre-questioning

O1, O2, O3, O4            : Stands for an observation or measurement of the subject in the

                            design treatment
        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. The effect of students’ gender on the students’ reading comprehension achievement.

3. The effect of treatments with pre-questioning, without pre-questioning and student’s

   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.



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
                                                              Students’ reading
                                                               comprehension
              Gender                                            achievements



3.4.3   Data Analysis Procedures

        To analyze the data, the writer followed the procedures:

1. Checking the students’ answers and give scores.

2. 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        -   Carefully ascertain that
                  time as the treatment may influence                no extraneous events
                  performance with regard to the outcome             occur
                  variable                                       -   Make sure that what
                                                                     happens to one group
                                                                 happens to other(s)
Selection bias    It may be the composition of the group     -   Make groups initially
                  rather than the treatment that accounts        equal-preferably        by
                  for performance with regard to the             random assignment
                  outcome variable                           -   Pretest to check initial
                                                                 performance of group(s)

Maturation        Improvements between one testing           -   Randomly assign groups
                  occasion and another may result from           to true experiment
                  routine changer that occur as time         -   Rule     out    maturation
                  passes                                         logically
                                                                 (e.g.development-
                                                                 ally implausible)
Instrumentation   Differences on various testing occasions   -   Carefully ascertain that
                  may occur because the data collection          no changes occur in the
                  process has somehow changed                    data collection process
                                                             -   If changes do occur, make
                                                                 them happen to both
                                                                 groups
Statistical       Groups selected on the basis of extreme    -   Don’t select on the basis
regression        score tend to shift toward the mean of         of extreme scores
                  the original group when re-tested          -   If you select extreme
                                                                 groups,     use     random
                                                                 assignment afterwards
Experimental      Differences on subsequent testing          -   Strive to prevent dropouts
mortality         occasions may occur because the            -   Analyze characteristics of
                  composition of the group has changed           dropouts
                                                             -   Use a matching strategy
                                                                 with subsequent random
                                                                 assignment
Pretesting        The experience of taking the pretest may   -   Do not use a pretest
                  sensitize subjects to perform better on    -   Use an unobtrusive pretest
                  the posttest                               -   Use random assignment to
                                                                 true experiment
Instability       Chance fluctuations in score arising       -   Use reliable tests
                  from unreliability may cause changes in    -   Use test to significance-
                  performance      on    various  testing        after random assignment,
                  occasions                                      if possible
Social-           Dynamic of the experimental situation      -   Make the experiment as
psychological     may set up alternative treatment that          natural and unobtrusive
threats           may account for differences                    as possible

Expectancy        Outcomes     may     occur     because     -   Discourage expectancies
effects           experimenter or subjects expected those    -   Foster opposite rather
                  outcomes                                       than           supportive
                                                                 expectancies

				
DOCUMENT INFO
Shared By:
Categories:
Tags: Coding
Stats:
views:21
posted:9/22/2012
language:English
pages:19
Description: The writer used a test to collect the data. After collecting the data, they were processed with the following procedures