A statistical methodology is proposed for the analysis of a latent concept which is fluctuating over time such as the perceived trust towards financial and political institutions. We conceive confidence as a mental unobservable feature of each person which is related to the observed time-varying and time-fixed covariates. We model the uncertainty in the responses through a hidden Markov model and we account for the longitudinal survey weights as well as for missing responses when survey data are available. We estimate the model parameters by a weighted log-likelihood which is maximized through the Expectation-Maximization algorithm in order to find hidden clusters of people with the same perceptions towards the institutions. We allocate each individual according to the Viterbi algorithm applied to the posterior probabilities. By considering the Polish society we find four hidden groups of Poles: discouraged, with no opinion, with selective trust and with fully trust towards institutions. We predict an increasing tendency to choose the institutions to support.

Pennoni, F., Genge, E. (2018). Predicting trends of institutional confidence through a hidden Markov model with survey weights and missing responses. In Book of Abstracts CFE-CMStatistics 2018 (pp.65-65).

Predicting trends of institutional confidence through a hidden Markov model with survey weights and missing responses

Pennoni, F;
2018

Abstract

A statistical methodology is proposed for the analysis of a latent concept which is fluctuating over time such as the perceived trust towards financial and political institutions. We conceive confidence as a mental unobservable feature of each person which is related to the observed time-varying and time-fixed covariates. We model the uncertainty in the responses through a hidden Markov model and we account for the longitudinal survey weights as well as for missing responses when survey data are available. We estimate the model parameters by a weighted log-likelihood which is maximized through the Expectation-Maximization algorithm in order to find hidden clusters of people with the same perceptions towards the institutions. We allocate each individual according to the Viterbi algorithm applied to the posterior probabilities. By considering the Polish society we find four hidden groups of Poles: discouraged, with no opinion, with selective trust and with fully trust towards institutions. We predict an increasing tendency to choose the institutions to support.
abstract + slide
Expectation-Maximization algorithm, Longitudinal data, Multivariate responses, Missing data, Trust- building policy discussion
English
International Conference of the ERCIM WG on Computational and Methodological Statistics 12th International Conference on Computational and Financial Econometrics University of Pisa, Italy
2018
Book of Abstracts CFE-CMStatistics 2018
978-9963222759
2018
65
65
http://cmstatistics.org/CMStatistics2018/
none
Pennoni, F., Genge, E. (2018). Predicting trends of institutional confidence through a hidden Markov model with survey weights and missing responses. In Book of Abstracts CFE-CMStatistics 2018 (pp.65-65).
File in questo prodotto:
Non ci sono file associati a questo prodotto.

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/10281/213445
Citazioni
  • Scopus ND
  • ???jsp.display-item.citation.isi??? ND
Social impact