That is, the latent variable (construct) can explain 
on average more than half of the variance of the 
indicators. 
Table 3: Results of the Average Variance Extracted Test 
with Smart PLS.
 
   AT  EE  FC  PE  PLO  SI  UH 
AT1 
0,782 0,434 0,266 0,561 0,562 0,332  0,332 
AT2 
0,768 0,517 0,341 0,588 0,510 0,252  0,301 
AT3 
0,719 0,488 0,288 0,496 0,465 0,234  0,354 
AT4 
0,801  0,811 0,409 0,548 0,429 0,269 0,287 
EE1 
0,480  0,796 0,267 0,483 0,368 0,302 0,219 
EE2 
0,565  0,806 0,221 0,607 0,545 0,151 0,267 
EE3 
0,391  0,703 0,217 0,406 0,360 0,332 0,330 
FC1 
0,273 0,182 0,845 0,236 0,148  0,283  0,190 
FC2 
0,412 0,323 0,727 0,447 0,336  0,251  0,193 
FC3 
0,298 0,185 0,851 0,237 0,152  0,269  0,194 
PE1 
0,623 0,626 0,356 0,856 0,631 0,283  0,147 
PE2 
0,648 0,526 0,354 0,854 0,536 0,348  0,262 
PE3 
0,501 0,491 0,304 0,787 0,579 0,354  0,251 
PLO1 
0,536 0,464 0,214 0,552 0,816 0,298 0,385 
PLO2 
0,572 0,542 0,262 0,625 0,881 0,301 0,264 
PLO3 
0,518 0,395 0,217 0,589 0,840 0,230 0,268 
PLO4 
0,568 0,518 0,285 0,618 0,890 0,286 0,242 
SI1 
0,261 0,209 0,244 0,348 0,250 0,810 0,161 
SI2 
0,325 0,314 0,306 0,338 0,303 0,865 0,198 
SI3 
0,289 0,226 0,289 0,321 0,227 0,833 0,162 
SI4 
0,319 0,334 0,284 0,322 0,316 0,877 0,220 
UH1 
0,359 0,245 0,168 0,180 0,223 0,167  0,810 
UH2 
0,319 0,324 0,300 0,305 0,361 0,233  0,777 
UH3 
0,296 0,253 0,100 0,132 0,214 0,119  0,776 
 
 
From table 3, we can conclude that all AVE 
values have a value greater than 0.5 so that it can be 
said that all variables meet the requirements for use 
and there are no problems in the AVE test. 
Disciminant Validity: This test is carried out 
with two stages of cross loading examination, namely 
cross loading between indicators and Fornell-
Lacker's cross loading (Subiyakto et al., 2014). 
Examination of the cross loading of each indicator is 
carried out by comparing the relationship between the 
indicator and its construct and other block constructs. 
If the correlation between the indicator and the 
construct is higher than the correlation with other 
block constructs, this indicates that the construct 
predicts the size of their block better than the other 
blocks (Afthanorhan & Asyraf, 2013; F. Hair Jr et al., 
2014; Hair et al., 2012; Henseler et al., 2009). 
Examination of Fornell-Lacker's cross loading 
was carried out by looking at the root AVE value 
which must be higher than the correlation between the 
construct and other constructs (Afthanorhan&Asyraf, 
2013; F. Hair Jr et al., 2014; Hair et al., 2012; 
Henseler et al., 2009; Subiyakto et al., 2014; 
Urbach&Ahlemann, 2010; Yamin& Kurniawan, 
2011). 
 
 
Figure 3: Analysis Model. 
3 RESULTS AND DISCUSSION 
3.1  A Myth of Ambu Hawuk 
Based on the results of ethnographic research, Ambu 
Hawuk's myth is obtained in the form of speech from 
several sources. The speech obtained is in the form of 
fragments that need to be combined into a whole unit. 
This causes the process of revitalizing myths needs to 
be reconstructed. 
Ambu Hawuk's myth is a myth that tells a sacred 
figure but is also believed to be a"hawuk"  or dark 
figure due to his black magic mastery.In a myth that 
is the basis of writing Ambu Hawuk poetry represents 
the contents of the myth that breaks all the 
assumptions about the figure of Ambu Hawuk as a 
figure  "Hawuk."  The myth of Ambu Hawuk 
represented by Ambu Hawuk 1 poem also reflects 
how the Sundanese human figure mastered the 
supernatural powers whose black or white 
classification lies in how to use them. Is it for good or 
badness to society? The figure of Ambu Hawuk who 
in the myth is told as a member of the kingdom who 
   Average Variance Extracted (AVE) 
AT  0,590 
EE  0,592 
FC  0,655 
PE  0,694 
PLO  0,735 
SI  0,717 
UH  0,621