Manfred te Grotenhuis

The reliability of a two-item scale: Pearson, Cronbach or Spearman-Brown?

Rob Eisinga, Manfred te Grotenhuis, Ben Pelzer Department of Social Science Research Methods and Department of Sociology, Radboud University Nijmegen, PO Box 9104, 6500 HE Nijmegen, The Netherlands October 8 2012 To obtain reliable measures researchers prefer multiple-item questionnaires rather than single-item tests. Multiple-item questionnaires may be costly however and time-consuming for participants to complete. They therefore frequently administer two-item measures, the reliability of which is commonly assessed by computing a reliability coefficient. There is some disagreement, however, what the most appropriate indicator of scale reliability is when a measure is composed of two items. The most frequently reported reliability statistic for multiple-item scales is Cronbach’s coefficient alpha and many researchers report this coefficient for their two-item measure

Education, Religiosity and Moral Attitudes: Explaining Cross-National Effect Differences

The aim of the present study is to answer three research questions on moral attitudes (i.e., attitudes concerning abortion, premarital and extramarital relations, and homosexual relations). Which parental and individual characteristics affect moral attitudes? Do the effects of parental and individual characteristics vary across countries? And, if so, can these effect differences on moral issues be explained by national characteristics? To answer these questions, we use the 1991 ISSP database containing relevant data of 16,604 inhabitants of 15 countries. Hypotheses are tested using multi-level analyses. We find that parental and individual religiosity, as well as individual educational attainment, have strong effects on moral attitudes. However, we observe considerable effect differences across countries, which is a rather new finding. Effects of individual religiosity on moral attitudes appear to be stronger in more religious countries and weaker in more secularized countries. Effects of individual education are stronger in more religiously heterogeneous countries and weaker in more religiously homogeneous countries. Finally, effects of individual education on moral attitudes are weaker in short-standing democracies than in long- standing ones.

Influential Cases in Multilevel Modeling A Methodological Comment

A large number of cross-national survey datasets have become available in recent decades. Consequently, scholars frequently apply multilevel models to test hypotheses on both the individual and the country level. However, no currently available cross-national survey project covers more than 54 countries (GESIS 2009). Multilevel modeling therefore runs the risk that higher-level slope estimates (and the substantial conclusions drawn from these estimates) are unreliable due to one or more influential cases (i.e., countries). This comment emphasizes the problem of influential cases and presents ways to detect and deal with them. To detect influential cases, one may use both graphic tools (e.g., scatter plots at the aggregate level) and numeric tools (e.g., diagnostic tests such as Cook’s D and DFBETAS). To illustrate the usefulness and necessity of these tools, we apply them to a study that was recently published in this journal (Ruiter and De Graaf 2006). Finally, we provide recommendations and tools to detect and handle influential cases, specifically in cross-sectional multilevel analyses.

Denomination, Religious Context, and Suicide: Neo-Durkheimian Multilevel Explanations Tested with Individual and Contextual Data

In Suicide, Durkheim found that involvement in religious communities is inversely related to suicide risk. In this article, two explanations for this relationship are examined. One is that religious networks provide support. The other is that religious communities prohibit suicide. To examine these hypotheses, individual‐level data on suicide in the Netherlands from 1936 to 1973 are used. The results show that with an increase in the proportion of religious persons in a municipality, the chances of committing suicide decrease for every denomination in that municipality, as well as among nonchurch members. Furthermore, along with the secularization of Dutch society, the impact of religious composition on suicide wanes. These results contradict the network‐support mechanism and confirm the notion that religious communities have a general protective effect against suicide.

Welfare States And Dimensions Of Social Capital: Cross-national Comparisons Of Social Contacts In European Countries

We set out to describe and explain differences in the amount of some dimensions of social capital within and between European societies. Social capital refers to a wide range of social phenomena; however, we focus on social contacts with family and friends. We derive hypotheses about cross-national differences in social capital from theories on the nature of welfare state regimes. We test these hypotheses with multi-level analyses on Eurobarometer data, collected in thirteen countries. We find significant variance across different countries. This variance is partly explained by individual characteristics: religious people and people living in medium-sized or rural towns have more social contacts. Moreover, we find quite differential effects of other individual characteristics on social contacts and no effects of political stances. Differences in the cross-national compositions in educational attainment and household size also account for the variance in social contacts. Finally, people living in social-democratic regimes turn out to have the smallest amount of social contacts, whereas people living in the Latin Rim have the largest amount. In between, we find people living in liberal, respectively, conservative-corporatist regimes. This explanation is opposed to the hypothesis that it is the difference in social security rates that causes differences in social capital.

Churches in Dutch: Causes of religious disaffiliation in the Netherlands 1937-1995

The Netherlands has become one of the most secular countries in the world. A vast majority of the Dutch people does not attend church regularly and more than half its population is not affiliated with any church at all. In this study we set out to test which individual and contextual characteristics affect religious disaffiliation. We deduced several hypotheses from theories on social integration and rationalization. To test these hypotheses we used retrospective data containing information on events that took place in the lives of our respondents since adolescence. These data were analysed using a discrete-time event history model. We found that the higher the level of rationalization in a certain year, the more likely people were to disaffiliate. This effect was particularly strong for young people. Moreover, by introducing rationalization in the model we found a number of spurious relationships that at first glance seemed to be causal. Not surprisingly, respondents were more likely to disaffiliate in cases where their partners were nonreligious. However, as respondents and their partners presumably are effected equally by rationalization, we cannot but conclude that the process of rationalization is mainly responsible for the process of religious disaffiliation that takes place in The Netherlands.

States as molders informal relations

ABSTRACT This article studies the impact of a range of state institutions on citizens’ contacts with family and close friends in 20 countries. Recent studies have shown large country level differences in social participation. We aim to explain these differences from an actor centered institutionalist perspective. We present two lines of reasoning. According to the first, a high level of social security crowds out social participation, as intimate networks are no longer needed as an economic safety net. The second line of reasoning proposes that corruption or a lack of civil rights drive citizens to seek refuge in their secure intimate contacts. In a comparative, multi-level design we focus on participation in the nuclear family, in the extended family, and with the best friend. We test the two lines of reasoning simultaneously on ISSP 2001 data. We find that states matter. State institutions are an important determinant of social participation. Our findings mainly confirm the second line of reasoning, whereas the crowding out thesis is only supported for contact with the extended family. Moreover, we find that the contextual effects are not similar across social groups: the poor are more strongly affected by the institutional design than the rich.

Interviewer BMI effects on under- and over-reporting of restrained eating: evidence from a national Dutch face-to-face survey and a postal follow-up

ObjectivesTo determine the effect of interviewer BMI on self-reported restrained eating in a face-to-face survey and to examine under- and over-reporting using the face-to face study and a postal follow-up.MethodsA sample of 1,212 Dutch adults was assigned to 98 interviewers with different BMI who administered an eating questionnaire. To further evaluate misreporting a mail follow-up was conducted among 504 participants. Data were analyzed using two-level hierarchical models.ResultsInterviewer BMI had a positive effect on restrained eating. Normal weight and pre-obese interviewers obtained valid responses, underweight interviewers stimulated under-reporting whereas obese interviewers triggered over-reporting.ConclusionIn face-to-face interviews self-reported dietary restraint is distorted by interviewer BMI. This result has implications for public health surveys, the more so given the expanding obesity epidemic.

The Non-uniqueness Property of the Intrinsic Estimator in APC Models

This article explores an important property of the intrinsic estimator that has received no attention in literature: the age, period, and cohort estimates of the intrinsic estimator are not unique but vary with the parameterization and reference categories chosen for these variables. We give a formal proof of the non-uniqueness property for effect coding and dummy variable coding. Using data on female mortality in the United States over the years 1960–1999, we show that the variation in the results obtained for different parameterizations and reference categories is substantial and leads to contradictory conclusions. We conclude that the non-uniqueness property is a new argument for not routinely applying the intrinsic estimator.

The intrinsic estimator, alternative estimates, and predictions of mortality trends: A comment on Masters, Hummer, Powers, Beck, Lin, and Finch

In this article, we discuss a study by Masters et al. (2014), published in Demography. Masters and associates estimated age, period, and cohort (APC) effects on U.S. mortality rates between 1959 and 2009 using the intrinsic estimator (IE). We first argue that before applying the IE, a grounded theoretical justification is needed for its fundamental constraint on minimum variance of the estimates. We next demonstrate IE’s high sensitivity to the type of dummy parameterization used to obtain the estimates. Finally, we discuss challenges in the interpretation of APC models. Our comments are not restricted to the article in question but pertain generally to any research that uses the IE.