Model-assisted multivariate analysis of social media engagement and public perception of governance in Nigeria.

Predominantly developing democracies are governed and perceived from the digitally structured public spheres and governed as democracies from social media platforms where the engagement and sentiment derived from the platforms shape the digitally manifested trust, satisfaction, and legitimacy of the citizens. However, the surveys, most often employed with great frequency, fail to grasp the intensity of the digital landscape. In the case of Nigeria, the study applied Advanced Multi-Variates Techniques Analytical Methodologies to enhance the estimation of the perceived governance and the panel social media metrics. In using structured random sampling of the 600 surveys of the citizens and the data analyzed from the social media accounts of the Government, the activated Principal Components Analysis, the Multiple Layers of Regression, and the General Regression Estimator were utilized to address governance of the citizens predicated on the engagement and sentiment. With the Principal Component Analysis, two latent factors were constructed: Civic Digital Engagement and Sentiment Tone with 78.4 percent of the variance in the digital behavior of the respondents. With GREG, the adjusted mean of the perceived governance was 54.86 and with the mean of the survey without data adjustment, the perceived mean was 52.84. Therefore, through the merged data sets, the improvement in the accuracy of the governance was inferred. In this case, the merged data set showed the sentiments of participants (β = 0.51, p < .001) and the tone (β=0.38, p < .001) of the sentiment. They also showed to an extent that the participants’ education and where a participant resides (rural/urban) affects the factors that improve the perception of governance. In this case, the multidimensional aspects of the Agenda-Setting and Framing theories were merged with Trust to explain how governance is and can be perceived. The findings confirm that incorporating model-assisted digital analytics in Nigeria's analytics frameworks can bolster transparency, participation and accountability and also improve policymaking.