Multinomial Logistic Regression as a Barometer of West Java People^s Level of Compliance in Implementing Health Protocols Sonny Harry Budiutomo Harmadi (a), Udjianna Sekteria Pasaribu (b*), Utriweni Mukhaiyar (b), Susi Setiyowati (b)
(a) Development Studies Department, Faculty of Creative Design and Digital Business, Institut Teknologi Sepuluh Nopember
(b) Statistics Research Group, Department of Mathematics, Faculty of Mathematics and Natural Science, Institut Teknologi Bandung, Indonesia
*corresponding author: udjianna[at]math.itb.ac.id
Abstract
West Java is the largest contributor to the addition of Covid-19. This incident has changed people^s behavior, so 3M^s health protocols must become the basis of daily life. The role of academics is very important as a basis for descriptions and submissions to the government. Therefore, research was conducted using a survey method to identify changes in people^s behavior when implementing 3M in West Java. Multinomial logistic regression is used to predict the dependent variable (y) with a nominal data type given 26 independent variables (x). The goal is to understand the compliance levels of 3M. It is linked to independent variables, namely (i) self-awareness, (ii) the effectiveness of 3M implementations, (iii) the level of community compliance in implementing these protocols, and (iv) information gathered from various sources. The obtained analysis shows that if the handshake intensity is reduced, the provincial government imposes strict rules and sanctions regarding the use of masks and requires handshakes, the people of West Java will be very obedient. The multinomial logistic regression equation for the people of West Java to comply with the 3M protocol is phi (x)=(exp(6,223-3,276x_5-2,884x_14+3,459x_18-2,698x_25 ))/(1+exp(6,223-3,276x_5-2,884x_14+3,459x_18-2,698x_25)), the probability of success.
Keywords: Covid-19, behavior change, health protocol, multinomial logistic regression
