The effect of science-technology-society approach-based worksheets on improving Indonesian students' scientific literacy http://www.ied.edu.hk/apfslt/v17

Scientific literacy consists of content (science as knowledge), process (science as a process of investigation and a way of thinking) and context (the interaction(s) amongst science, technology, and society). The results of the normalized gain scores (n-Gain) for each domain of the grade 7 students' scientific literacy levels are displayed in Table 4.

Table 4. The results of n-Gain and independent samples t-test

Group	*Mean*			Criteria	Uji-t n-Gain Sig. (2- tailed)
Group	Pre test (x̄ ± sd)	Post test (x̄ ± sd)	n-Gain (x̄ ± sd)	Criteria	Uji-t n-Gain Sig. (2- tailed)
Control (n=32)	44.4 ± 8.41	67.6 ± 6.62	0.48 ± 0.13	Medium	Significantly different (pPretes=0,017, pPost tes=0,000, pn-Gain=0,000)
Experiment (n=32)	49.4 ± 7.95	77.3 ± 7.28	0.66 ± 0.14	Medium

As seen in Table 4, n-Gain value of the experimental group was found to be 0.66, meaning a significant improvement of grade 7 students' scientific literacy levels from pretest to posttest, whilst that for the control group was calculated to be 0.48, meaning an improvement for low rank students in scientific literacy level. Therefore, improvements significantly occurred in the experimental group. Based The value of Sig. (2-tailed) <0.05 rejected to a significant difference between the experimental and control groups’ n-Gain scientific literacy levels. An increase in each student’s n-Gainis shown in Figure 2. N-Gain scores of grade 7 students' scientific literacy levels on domains of the content, process, and context of science illuminates that the experimental group had a higher n-Gain average, which fell into the moderate category.

Figure 2.N-Gain scores of grade 7 students' scientific literacy levels

Finally, a power analysis to determine the minimum effect size of the quasi-experimental research design was conducted. For analysis of grade 7 students’ skills in pre-test and post-test, a total of the sample of the study was assumed to be N= 64. That is, there were two groups and two assessment times (i.e., pre-test and post-test). Effect size of the least analyse fort he experimental group was found to be p.η² = 0.767.