对调研数据进行量化分析的两种新方法,希望对大家能有启发
本帖隐藏的内容
ABSTRACT
In this paper we propose and evaluate two new methods for the quantifi cation
of business surveys concerning the qualitative assessment of the state of the
economy. The fi rst is a nonparametric method based on the spectral envelope,
originally proposed by Stoffer, Tyler and McDougall (Spectral analysis for
categorical time series: scaling and the spectral envelope, Biometrika 80: 611–
622) to the multivariate time series of the counts in each response category.
Secondly, we fi t by maximum likelihood a cumulative logit unobserved
components models featuring a common cycle. The conditional mean of the
cycle, which can be evaluated by importance sampling, offers the required
quantifi cation. We assess the validity of the two methods by comparing
the results with a standard quantifi cation based on the balance of opinions and
with a quantitative economic indicator.