Multilevel Regression

The objective of this paper is to describe cross-country similarities and differences in national pride and to explain national pride variations on the individual and country levels. The analysis in this paper is applied to different measures of national pride, with some of them being relatively complex cognitively and the others more elementary. The paper presents the results of cross-country comparison of national pride based on empirical evidence from the ISSP-2003 database which included data from 45993 respondents from 36 countries and regions. The survey participants estimated their overall level of national pride by responding to the direct one-item question and, separately, they estimated pride of each of ten specific achievements of their countries in various domains. Factor analysis of these ten items yielded two dimensions of domain-based national pride, one of them being the factor of general pride of various country achievements and the other reflects the inverse relations between the prides of elitist and mass achievements of the nation.

The multilevel regression models estimated for the three indicators of national pride confirm the feasibility of dividing these indicators into cognitively processed and normatively imposed national pride. Cognitively processed national pride measured by the domain-based estimates have been affected by objective country achievements and by the level of standards which the achievements are compared against. The normatively imposed national pride measured by direct one-item question has been influenced by the country level of religiosity that indicates the individual willingness to accept normative messages from the state uncritically.

Rational national pride requires some objective grounds to believe in a nation’s perfection, and normative national pride is not so strongly related to objective achievements and therefore can be more easily manipulated. The practical implication of this difference stems from the fact that in their search for objectively grounded national pride people would be eager to foster country achievements and their maintenance of normatively imposed pride requires in many cases just reliably protected wishful thinking.

Bu kitap, araştırmacılara çok düzeyli regresyon modellemelerini ve HLM programı aracılığıyla analizin nasıl gerçekleştirileceğini açıklamak için hazırlanmıştır. Dört bölümden oluşan kitabın ilk bölümünde çok düzeyli regresyon modellemelerinin istatistiksel temellerine değinilmeye çalışılmıştır. İki, üç ve dördüncü bölümlerde ise tamamen örnek bir araştırma durumu üzerinden uygulamalara yer verilmiştir. Bu bölümlerde, öncelikle örnek olabilecek bir araştırma durumu açıklanmış, sonra bu araştırma durumuna ait verilerin HLM programına nasıl aktarılacağı ve analizin gerçekleştiriliş biçimi görsellerle detaylı olarak açıklanmıştır. Araştırmacılar için en önemli sorunsallardan birisi yaptığı analizlerin nasıl raporlanacağı olduğundan, raporlamaya ilişkin birer tablo ve yorum örneğine de uygulamaların yapıldığı bölümlerde yer verilmiştir. Kitapta teorik temellerden çok analizin uygulanabilmesi ön planda tutulmuştur.

Stress tests have recently become one of the usual procedures to assess the resilience of the EU banking systems against economic distress. Their results have also influenced and conditioned the last European banking sector restructuring and recapitalization process.

The aim of this paper is to predict the core tier 1 ratio assessed by the 2011 European stress tests by means of a multilevel regression model. Financial ratios, sovereign debt exposures and macroeconomic indicators are considered as explanatory variables. The sample consists of the financial institutions from the so-called PIIGS countries (Portugal, Italy, Ireland, Greece and Spain). The results show that a CAMEL model has good predictive capability when country and scenario dummies and their interaction are included.

Introduction: Nepal won Millennium Development Goal Award on Maternal Health as the Maternal Mortality declined by 50% between 1996 and 2006 and by 80% between 1990 and 2010. However, maternal mortality in Nepal are estimated from Surveys using direct and indirect techniques amid the absence of robust vital registration system. Thus, maternal mortality estimates are not available at regional and district levels.

Methods: This paper estimated the regional and district level maternal mortality ratio for 2011 using multilevel regression model coefficients proposed by UN and World Bank for the first time in Nepal. Proportion of Maternal Deaths were estimated using Gross Domestic Product, General Fertility Rate and delivery by Skilled Birth Attendants. The maternal mortality ratio for five development regions and 75 districts were then estimated using the 2011 Census Reports.

Results: The maternal mortality ratio at 2011 was estimated as 155 deaths per 100,000 live births and was consistent with the previous values estimated between 1990 and 2010. Maternal mortality at Western and Central development regions were below while Eastern, Mid-Western and Far-Western Regions were above the national estimate. Twenty-two districts had it below while 53 had it above the national estimate. Kalikot and Kaski districts had highest and lowest maternal mortality ratio
in 2011.

Conclusion: Maternal mortality declined in the country because of rapid fertility decline and increased contraceptive use and decreased anemia among pregnant women and obstetrics/abortion facilities in all districts of the country after legalization of abortion. Improvement in women’s education, empowerment, wealth and living standard were also responsible for declining maternal mortality. Nepal now needs a robust vital registration system to “explain” the reasons of this decline and further devise and implement new plans, policies and programs to reduce it further at sub-national and local levels.