Technique of Analyzing the Data

F. Technique of Analyzing the Data

1. Trying Out the Instrument

a. Validity

The instrument on the research is defined as a tool for collecting the data or information. Before using the instrument, there are two kinds of validities done, namely internal and external validity. The internal validity is done when the validators check the analysis sheets.

According to Sudjana (2005:12), to know the real result and process of the students’ learning accurately depend on the quality of instrument beyond the procedures. However, instrument is valid or contains a very high validity only if it can measure what it should be measure (Anggoro, et al., 2011:5.28; Andriani, et al., 2013:5.27; Sutomo, 1985:48).

In the research, the researcher tried out the instrument internally because it has rational validity. The instrument has rational or internal validity only if the criteria that is found on the instrument itself reflect to things what will be measured

(Sugiyono, 2011:123). Here, the instrument’ validators checked the analysis sheet to make sure that the instrument is appropriate to be used or not.

b. Reliability

Reliability is defined as the constancy of the instrument to get the data of the research. Anggoro, et al., (2011:5.31) affirms that reliability means the accuratenes of the instrument for measuring. Follow on, Sudjana (2005:16) asserts that if the instrument has constancy to measure what it should be measured, of course it is reliable. The constancy itself is the situation which is the instrument gives the same result even the instrument itself is used many times in different occasion.

In the research, the researcher examined the reliability of the instrument internally because it also has rational consistency. As Sugiyono (2011:122) writes that a reliable instrument is not definitely valid, but a valid instrument is trusted reliable. So, the researcher sought the internal validity of the instrument to know its reliability.

2. Analyzing the Data

a. The Students’ Ability in Writing Skill

In conducting the research, the researcher used the recount text as the topic of teaching material. So, to know the studen ts’ ability in writing skill, the researcher used the test. The result of the research had been administered by using scoring profile as suggested by Weir (1990) in Weigle (2002:117) to assess the students through a personal writing.

In giving the score, the researcher focused on several criteria as follows.

Table 4

ANALYTICAL EVALUATION RUBRIC of PERSONAL WRITING (Suggested by Weir (1990) in Weigle (2002:117))

Point Score

Information

A Relevance and Adequacy of content

0 The answer bears almost no relation to the task set. Totally

inadequate answer.

1 Answer of limited relevance to the task set. Possibly major gaps in

treatment of topic and / or pointless repetition.

2 For the most part answers the tasks set, though there may be some

gaps or redundant information.

3 Relevant and adequate answer to take set.

Continued of Table 4…

B Compositional organization

0 No apparent organization of content.

1 Very little organization of content. Underlying structure not

sufficiently controlled.

2 Some organizational skills in evidence, but not adequate controlled.

3 Overall shape and internal pattern clear. Organizational skills

adequately controlled.

C Cohesion

0 Cohesion almost totally absent. Writing so fragmentary that comprehension of the intended communication is virtually impossible.

1 Unsatisfactory cohesion may cause difficulty in comprehension of

most of the intended communication.

2 For the most part satisfactory cohesion although occasional deficiencies may mean that certain parts of the communication are not always effective.

3 Satisfactory use of cohesion resulting in effective communication.

D Adequacy of Vocabulary for purpose

0 Vocabulary inadequate even for the most basic parts of the intended

communication.

1 Frequent inadequacies in vocabulary for the task. Perhaps frequent

lexical inappropriate and / or repetition.

Continued of Table 4…

2 Some inadequacies in vocabulary for the task. Perhaps some lexical

inappropriate and / or circumlocution.

3 Almost no inadequacies in vocabulary for the task. Only rare

inappropriate and /or circumlocution.

E Grammar

0 Almost all grammatical patterns inaccurate.

1 Frequent grammatical inaccuracies.

2 Some grammatical inaccuracies.

3 Almost no grammatical inaccuracies.

F Mechanical Accuracy I (Punctuation)

0 Ignorance of conventions of punctuation.

1 Low standard of accuracy in punctuation.

2 Some inaccuracies in punctuation.

3 Almost no inaccuracies in punctuation.

G Mechanical Accuracy II (Spelling)

0 Almost all spelling inaccurate.

1 Low standard of accuracy in spelling.

2 Some inaccuracies in spelling.

3 Almost no inaccuracies in spelling. Obtained Score = A + B + C + D + E + F + G Maximum Score = 7 x 3 = 21 Constant Number = 100

Then the students’ mark is calculated through formula as suggested by Arifin, (2012:140) like as.

Next, the researcher classified of the students’ proficiency by using some criteria as follows.

Scales:

100 = Excellent 85-99

= Very good 75-84

= Good 65-74

= Good enough 55-64

= Enough 45-54

= Bad 0-44

= Very bad

b) Mean Score and Standard Deviation

To know the mean score and standard deviation of the students’ test result even in experimental class and control class, the researcher counted it by using the formula as suggested by Sudjana (2005:109) as follows.

∑X

Which:

X = the mean of the data ∑X

= the sum of all scores N

= the numbers of sample

To get the standard deviation, the researcher used the formula from Sudjana (2005:114):

S=

Which:

= the mean of the data S

= the standard deviation of the sample

X = the score of item N

= the numbers of sample

c) Normality

In seeking the normality in functions to know whether the sample comes from population which had the normal distribution or not, the researcher used Liliefors’ formula as Herhyanto, et al., (2014:8.17) suggest like as.

(1) Make the work table as given next. (2) Input the data or score in the table. (3) Find the Z score by using the formula:

Xi− X

S Which.

Z i = the standard score

X i = the score of i S = the standard deviation

X = the mean score

(4) Determine the F (z i ) through Table Standard Normal based on the Z score. 𝐹 𝑐𝑢𝑚𝑚𝑢𝑙𝑎𝑡𝑖𝑣𝑒

(5) Determine the S (z) by using the formula: S(z) =

(6) Count the quarrel between F(z) and S(z) by using the formula. L c = │F(z) – S(z)│(decide the highest one as Liliefors count (L 0 ).

(7) Determine the score of Liliefors table (L t ) with 𝛼 = 0,05. (8) Compare Liliefors count and Liliefors table, then take the conclusion like as.

(a) If L 0 < Lt, so the data is normal and the population has the normal distribution. (b) If L 0 > Lt, so the data is not normal and the population has not normal distribution.

Table 5

WORK TABLE for LILIEFORS FORMULA

F (Z i )

S (z i )

F(z i )-S(z i )

│F(z i )-S(z i )│

L count

L table

Conclusion

d) Homogeneity

To seek the homogeneity of the sample, the researcher sought it by using Harley formula as suggested by Irianto, et al., (2007:276) through several steps as follows.

(1) Assure that population has normal distribution. (2) Count the mean score of the experimental class and control class. (3) Count the variance of the both classes through formula like as.

(4) Determine F count with formula.

(5) Determine F table through formula.

𝐹 𝑡𝑎𝑏𝑙𝑒 = 𝐹𝛼 (𝑛 ℎ𝑖𝑔ℎ𝑒𝑠𝑡 𝑣𝑎𝑟𝑖𝑎𝑛𝑐𝑒 − 1, 𝑛 𝑙𝑜𝑤𝑒𝑠𝑡 𝑣𝑎𝑟𝑖𝑎𝑛𝑐𝑒 − 1), then confirm it on the table of F distribution critics value.

(6) Compare F count and F table and get decision on it as follows. If F count ≤F table , so the both samples are homogeny. If F count >F table, so the both samples are not homogeny.

e) Hypotheses Testing

The result of the both groups (experimental and control class) was be used by the researcher to tests the hypotheses such as. Ha : There is a significant effect of Modelled Writing on students’ ability in writing skill at the eighth grade of SMP Negeri 1 Ulugawo in 2015/2016.

H 0 : There is no any significant effect of Modelled Writing on students’ ability in writing skill at the eighth grade of SMP Negeri 1 Ulugawo in 2015/2016. Recognizing the data of the posttest result showed the normal distribution and homogeny, the researcher tested the hypothesis by using parametric statistic formula. In the research, the researcher used the t test of separated variant formula as suggested by Sugiyono (2011:197) as follows.

Which: t

=t count

X 1 = the mean of experimental group

X 2 = the mean of control group

62

n 1 = total of experimental group n 2 = total of control group

𝑆 2 1 = variants of experimental group 𝑆 2

2 = variants of control group In the overall activity, the researcher confirmed the t table which the level of

significance 0.05 and df = n 1 +n 2 -2. Lastly, in testing criteria; since t count was higher than t table , so H a was accepted.