CHAPTER IV: RESEARCH FINDING AND DISCUSSION
2. Validity and Reliability of Test
First, the test was tried out to the non-respondent class (X IPS 2) on 13-20of May 2022. The participant was 34 students. The test was a completion test with 20 fill-in-the-blanks of simple past tense and a subjective test with some guided questions of writing personal recount text. This try-out was recently used to check the difficulty and effectiveness level of the test.
a. Validity of Test 1) Face Validity
Mostly, the students of X IPS 2 agreed that the questions of completion and the subjective test seemed quite clear for them.
(See Appendix 3)
According to the English teacher, “some questions were relevant to the basic competence of simple past tense and writing a recount text. Afterward, she agreed that the fill-in- the-blanks method was efficiently used to find how is the students‟ understanding of simple past tense deeply. At the same time, the guided writing method seemed like the appropriate measurement for capturing a good recount text”.
(See Appendix 4)
Hence, it showed that both of the tests have face validity.
2) Content Validity
According to the English teacher, both of the tests were relevant to the basic competence of simple past tense and writing a recount text.
It could be seen from the lesson plan (See Appendix 5 & 6) and the English syllabus of the 10th grade (See Appendix 7).
Both of them seemed quite suitable with the blue print of the tests. (See Appendix 8)
It means that the contents of the test were correctly related to the learning material of simple past tense and writing a recount which have been taught before by the English teacher.
Thus, it showed that both of the tests have a content validity.
3) Item Validity
There is a criterion to know the validity of items. In this research, the item of the test is considered as valid if the correlation coefficient is at least the same as the r table at 0.05 for df (N-2) which is 0.3388.
Table 4.2
r-table for df (N-2)/ N = 34, df (34-2) = 32
df = (N-2)
Tingkat signifikansi untuk uji dua arah 0.1 0.05 0.02 0.01 0.001 32 0.2869 0.3388 0.3972 0.4357 0.5392
a) Item validity of simple past tense test
The first simple past tense test was held on Friday, the 13th of May 2022 at X IPS 2 as the non-respondent class. This test was completion test with 20 items of fill-in- the-blank questions. The score of each question was accumulated by the researcher. (See Appendix 9) Then, it would be calculated by Pearson’s correlation product moment (IBM SPSS Statistics 22 Version).
Thus, the result of the item validity which was calculated by using Pearson’s correlation product moment (IBM SPSS Statistics 22 Version) as follows:
(or See Appendix 10)
Table 4.3
The First Item Correlation of Simple Past Tense Test Item r value r table (0.05)
S1 0.370 0.338 Valid
S2 0.370 0.338 Valid
S3 0.370 0.338 Valid
S4 0.460 0.338 Valid
S5 0.570 0.338 Valid
S6 0.364 0.338 Valid
S7 0.438 0.338 Valid
S8 0.541 0.338 Valid
S9 0.479 0.338 Valid
S10 0.459 0.338 Valid
S11 0.343 0.338 Valid
S12 0.255 0.338 Invalid
S13 0.126 0.338 Invalid
S14 0.238 0.338 Invalid
S15 0.247 0.338 Invalid
Item r value r table (0.05)
S16 0.605 0.338 Valid
S17 0.774 0.338 Valid
S18 0.751 0.338 Valid
S19 0.774 0.338 Valid
S20 0.774 0.338 Valid
It could be seen from the table above, there were 20 items or questions of simple past tense test. However, there were 4 invalid items on the first test of simple past tense mastery. The items were the number of twelve until fifteen.
Those were invalid because the r value was less than r table (r
value < 0.338).
In those items, students were asked to choose and change the suitable “to be” which was “(is/are), or (is not/are not)” became the simple past form in a sentence.
(See Appendix 11)
However, some students said that they were being confused because some questions seemed similar with a “to be” of present tense. It was proven by seeing students‟
answer of those questions.
Hence, the researcher found some students did not change the words in column became the simple past form.
But, it was still answered by using the same word which was “is or are, and is not or are not”. (See Appendix 12)
From that first result, the researcher decided to give second tried-out that was held on Friday, the 20th of May 2022 at the non-respondent class too (X IPS 2). This second tried-out was completion test with 20 fill-in-the-blank questions. Afterward, the score of each question was accumulated again by the researcher (See Appendix 13) and would be calculated by using Pearson’s Correlation Product Moment (IBM SPSS Statistics 22 Version).
Thus, the result of the second item validity which was calculated by using Pearson’s correlation product moment (IBM SPSS Statistics 22 Version) as follows:
(or See Appendix 14)
Table 4.4
The Second Item Correlation of Simple Past Tense Test Item r value r table (0.05)
S1 0.775 0.338 Valid
S2 0.870 0.338 Valid
S3 0.524 0.338 Valid
S4 0.524 0.338 Valid
S5 0.807 0.338 Valid
S6 0.807 0.338 Valid
S7 0.662 0.338 Valid
S8 0.767 0.338 Valid
S9 0.535 0.338 Valid
S10 0.700 0.338 Valid
S11 0.524 0.338 Valid
S12 0.807 0.338 Valid
S13 0.662 0.338 Valid
S14 0.700 0.338 Valid
S15 0.767 0.338 Valid
Item r value r table (0.05)
S16 0.775 0.338 Valid
S17 0.819 0.338 Valid
S18 0.870 0.338 Valid
S19 0.819 0.338 Valid
S20 0.819 0.338 Valid
It could be seen from the table above, those invalid questions became valid at the second validity of simple past tense test. Those items were valid because r value was higher than r table. In this second test, the researcher revised 4 invalid questions which were the number of 12 – 15.
At the first test, those questions were about the changing of “to be” in simple past tense. However, it still brought up students‟ confusion between a “to be” of simple present and simple past tense. Hence, the researcher tried to make them clear by changing the words “is/are” and “is not/are not” became this word “to be”. (See Appendix 15)
Hence, students‟ value of those questions got better than before. (See Appendix 13) It was proven by seeing students‟ answer where mostly of them gave a correct answer to the questions. (See Appendix 16) They said that those questions were easy to understand than the first test before.
Thus, the second tried-out of simple past tense test was valid and would be used at X MIPA 2 as the respondent class in this research.
b) Item validity of writing personal recount text
The writing test of personal recount text was held on Wednesday, the 18th of May 2022 at X IPS 2 as the non- respondent class. This test was subjective test with some guided questions to make paragraphs of personal recount text. The score of each question was accumulated by the researcher. (See Appendix 17) Then, it would be calculated by Pearson’s Correlation Product Moment (IBM SPSS Statistics 22 Version).
Here, the result of item validity of writing a recount which was calculated by using Pearson’s Correlation Product Moment (IBM SPSS Statistics 22 Version) as follows:
(or See Appendix 18)
Table 4.5
Item Correlation of Writing Personal Recount Text Item r value r table (0.05)
Q1 0.924 0.338 Valid
Q2 0.935 0.338 Valid
Q3 0.923 0.338 Valid
It could be seen from the table above that all the questions of writing test were categorized as valid. It was
that all the r value of those questions was higher than r table. The result of students‟ writing test was also better than researcher‟s expectation. From those guided questions, most students could write a good personal recount text as well. (See Appendix 19)
Students said that a guided question and researcher‟s instruction were very clear to help them write.
It also concluded that this writing test could be directly used at X MIPA 2 as the respondent class later.
b. Reliability of Test
a) Reliability of Simple Past Tense Test
Besides that, the researcher used Cronbach’s Alpha by IBM SPSS Statistics 22 Version to check the reliability of the first simple past test. In the first test, there were 16 questions which were categorized as a valid item. For that result, the researcher only used the valid items in this first reliability.
The test is considered as reliable if the value of Cronbach’s Alpha is higher than 0.60. Hence, the result of reliability test of first simple past tense test which was calculated by using Cronbach’s Alpha (IBM SPSS Statistics 22 Version) as follows:
Table 4.6
The Reliability Statistics of First Simple Past Tense Test
Based on the table above, the reliability statistics of first simple past tense test showed that the value of Cronbach’s Alpha was 0.740 with 16 items. It was higher than 0.60 (0.740>0.60). The level of reliability showed that the value was between >0.60 – 0.80 (which was 0.740). It means that the 16 items of first simple past tense test were reliable.
Afterward, the researcher also used Cronbach’s Alpha by IBM SPSS Statistics 22 version to check the reliability of the second simple past tense test. In the second test, all the items were categorized as a valid instrument of simple past tense (20 questions). For that result, all the 20 items were used by the researcher in this second reliability of simple past tense test.
Hence, the result of reliability test of second simple past tense test which was calculated by using Cronbach’s Alpha (IBM SPSS Statistics 22 Version) as follows:
Table 4.7
The Reliability Statistics of Second Simple Past Tense Test
Cronbach's
Alpha N of Items
,740 16
Reliability Statistics
Cronbach's
Alpha N of Items
,919 20
Reliability Statistics
Based on the table above, the reliability statistics of second simple past tense test showed that the value of Cronbach’s Alpha was 0.919 with 20 items. It was higher than 0.60 (0.919>0.60). The level of reliability showed that the value was between >0.80 – 1.00 (which was 0.919). It means that the 20 items of second simple past tense test were very reliable.
b) Reliability of Writing Personal Recount Text
To check the reliability of writing personal recount text, the researcher also used Cronbach’s Alpha by IBM SPSS Statistics 22 Version. In the writing test, all the guided questions were categorized as a valid instrument. From that result, all the guided questions were used in this reliability of writing test.
Hence, the result of reliability of writing test which was calculated by Cronbach’s Alpha (IBM SPSS Statistics 22 Version) as follows:
Table 4.8
The Reliability Statistics of Writing Personal Recount Test
Based on the table above, the reliability statistics of writing personal recount test showed that the value of Cronbach’s Alpha was 0.918 with 3 items. It was higher than 0.60 (0.918>0.60). The level of reliability showed that the value was
Cronbach's Alpha N of Items
,918 3
Reliability Statistics
between >0.80 – 1.00 (which was 0.918). It means that the 3 items of writing personal recount test were very reliable.
c. Practicality
Based on the observation and experimentation at the non- respondent class (X IPS 2), the researcher used the practicality checklist as follows:
Table 4.9
The Practicality Checklist
Checklist Details Is the instrument
expensive?
Both of the tests are in the minimum budget.
Is the instrument easy to administer?
The administrative details clearly established before the test.
Does It have a specific and effective scoring procedure?
Both of the tests were valid and reliable.
Since it was found that the instruments were easy to administer, score, analyze and interpret, both of the instruments were practical.