4.4. The influence of socio-economic characteristics and
supporting factors for HSW reduction
The results of the binary logistic regression on the community
sorting activity are shown in Table 6. Based on the result, the influencing
factors for the HSW sorting activity were age of respondent,
the level of knowledge, the presence of an environmental cadre,
and the waste bank availability because the significant values were
0.000; 0.000; 0.064; and 0.000, respectively. The significant values
were less than 0.1 (10%), indicating that the variables highly
influence the sorting activity.
Based on Eq. (2), the logistic regression model on sorting activity
was as follows:
g (x) = −2.766 + 2.035 age (1) + 0.968 age (2) + 1.113 age (3)
+0.768 level of knowledge (1) + 1.943 level of knowledge (2)
+0.588 environmental cadre (1) + 1.518 waste bank (1)
The probability of sorting HSW by respondent with the age of
36–50 years old was eight times more than it with the age of less
than 35 years old. The probability of sorting HSW by respondent
with the age of 51–65 years old was three times more than it with
the age of less than 35 years old. The probability of sorting HSW by
respondent with the age of more than 65 years old was three times
more than it with the age of less than 35 years old.
The probability of sorting HSW by respondent with level of
knowledge “sufficient knowledge” was two times more than it with
the level of knowledge “low level of knowledge”. Additionally, the
probability of a respondent with the knowledge level indicated
as “high level of knowledge” was seven times more likely to sort
HSW than a respondent with the knowledge level of “low level of
knowledge”.
While the probability of sorting was two times higher when an
environmental cadre was available than when there was no environmental
cadre. In addition, the probability of sorting was five
times higher when, a waste bank was available than when there
was no waste bank.
The fitted model of sorting activity was used to estimate the
probability on sorting activity. For example, respondent of 40 years
old (=1) with high level of knowledge (=1), there were an environmental
cadre (=1) and a waste bank in his area (=1). So, the
probability of sorting activity was as follows:
g (x) = −2.766 + 2.035 (1) + 0.968 (0) + 1.113 (0) + 0.768 (0)
+1.943 (1) + 0.588 (1) + 1.518 (1)
g (x) = 3.318
(x) = e[3.318]
1 + e[3.318]
= 0.965 ≈ 0.97 = 97%
On the other hand, if there were no an environmental cadre (=0)
and a waste bank (=0), the probability of sorting was as follows:
g (x) = −2.766 + 2.035 (1) + 0.968 (0) + 1.113 (0) + 0.768 (0)
+1.943 (1) + 0.588 (0) + 1.518 (0)
g (x) = 1.212
(x) = e[1.212]
1 + e[1.212]
= 0.77 = 77%
The estimated logistic probability on respondent’s sorting activity
was 97%, if he was 40 years old with high level of knowledge, and
there were an environmental cadre and a waste bank in his area.
When there were not an environmental cadre and a waste bank,
the estimated logistic probability on respondent’s sorting activity
was 77%.
Table 7 reports the results from the binary logistic regression
on creating unique handcrafted goods from recyclable waste. The
influencing factors for creating unique handcrafted goods from
recyclable waste were the availability of environmental cadre
and waste bank, because the significant values were 0.066 and
0.011, respectively. The significant values are less than 0.1 (10%),
indicating that the variables exert substantial control over the
recycling activity. The socio-economic characteristics had no significant
influence in the respondent’s activity in creating unique
handcrafted goods from recyclable waste. The supporting factors
influenced the respondent’s activity in creating unique handcrafted