Shrinkage inverse regression estimation for model-free variable selection

Summary.  The family of inverse regression estimators that was recently proposed by Cook and Ni has proven effective in dimension reduction by transforming the high dimensional predictor vector to its low dimensional projections. We propose a general shrinkage estimation strategy for the entire inverse regression estimation family that is capable of simultaneous dimension reduction and variable selection. We demonstrate that the new estimators achieve consistency in variable selection without requiring any traditional model, meanwhile retaining the root n estimation consistency of the dimension reduction basis. We also show the effectiveness of the new estimators through both simulation and real data analysis.     

Assessing the Playing Field: A Prospective Longitudinal Study of Internet Sports Gambling Behavior

Internet gambling is growing rapidly, as is concern about its possible effect on the public’s health. This paper reports the results of the first prospective longitudinal study of actual Internet sports gambling behavior during eight study months. Data include recorded fixed-odds bets on the outcome of sporting contests and live-action bets on the outcome of events within contests for 40,499 Internet sports gambling service subscribers who enrolled during February 2005. We tracked the following primary gambling behaviors: daily totals of the number of bets made, money bet, and money won. We transformed these variables into measures of gambling involvement. We analyzed behavior for both fixed-odds and live-action bets. The median betting behavior of the 39,719 fixed-odds bettors was to place 2.5 bets of €4 (approximately $5.3 US) every fourth day during the median 4 months from first to last bet. This typical pattern incurred a loss of 29% of the amount wagered. The median betting behavior of the 24,794 live-action bettors was to place 2.8 wagers of €4 every fourth day during the median duration of 6 weeks at a loss of 18% of the amount wagered. We also examined the behavior of empirically determined groups of heavily involved bettors whose activity exceeded that of 99% of the sample.     

Staying safe while consuming alcohol: a qualitative study of the protective strategies and informational needs of college freshmen

OBJECTIVE: In this qualitative study, the authors examined how students attempt to minimize harm to themselves and others when drinking. PARTICIPANTS: The authors recruited freshmen at a large, mid-Atlantic US public university during the fall semester of 2005 to participate in 8 focus groups. METHODS: The moderator's guide was developed through an iterative process that included input from experts and pilot testing. The researchers audiotaped focus group conversations, transcribed them, and subjected them to an interrater reliability check. Analysis was based on the framework of Information-Motivation-Behavioral Skills Model and a phenomenological approach. RESULTS: College students have a repertoire of coping strategies they use in an attempt to safeguard themselves and their friends from harm when drinking. Strategies encompass planning a safe context for drinking, using safety measures to minimize harm when drinking, and taking care of someone who has consumed too much alcohol. CONCLUSIONS: A harm-reduction focus that acknowledges and builds on existing protective strategies may be a promising avenue for alcohol interventions.     

College and university compliance with a required meningococcal vaccination law

OBJECTIVE: Maryland became the first state to pass a vaccination law requiring college and university students living on campus to obtain a meningococcal vaccination or to sign a waiver refusing vaccination because college students are at increased risk for disease. The authors sought to identify how Maryland colleges addressed the law and determine whether schools were in full compliance. PARTICIPANTS: The authors surveyed 32 college/university administrators via a self-administered questionnaire. METHODS: The authors calculated vaccination and waiver rates and assessed compliance with the law overall and with specific law components. RESULTS: Among 28 participating schools, annual vaccination rates and waiver rates among students during 2000-2004 ranged from 66%-76% and 12%-17%, respectively. Two (7%) schools were compliant with all components of the law. CONCLUSIONS: Mandatory vaccination laws do not ensure compliance at the college and university level. Mandatory reporting, increased education, and collaboration between colleges and universities and public health agencies are needed.     

Discretionary Use of Computers: An Empirically Derived Explanatory Model*

This study identifies variables that explain variations in computer use. Factor analysis of data gathered from 422 business administration faculty reveal eight themes among the independent variables. Seven of these factors significantly discriminate among non, minimal, and high users. These results suggest an empirical model for future studies and provide policy insights for decision makers seeking to stimulate computer usage. Additionally, the study shows that there is much commonality between the previously separate organizational innovation and information systems implementation streams. It also demonstrates that motivation theory may be a valid framework within which to study computer use.     

The effect of media sexism on women’s political ambition: evidence from a worldwide study

This paper presents results from one of the first global studies on the relationship between media sexism and the share of candidates for the lower chamber of national parliaments who are women. Data on media sexism come from the Global Media Monitoring Project, the most reliable worldwide source for media coverage from a gender perspective. The data on share of female political candidates come from the Inter-Parliamentary Union. The results show that that there is, even when controlling for the number of women in parliament, electoral system, gender quotas, level of human development, level of women’s rights, freedom from corruption, and media access, a significant relationship between media sexism—measured as (i) the share of all news subjects that are women and (ii) the share of all news subjects portrayed in the function of experts who are women—and the share of women candidates for parliament: the higher the level of media sexism, the lower the share of women candidates. The theory discussed zooms in on a bystander effect: We hypothesize that sexist portrayals of women in the media stifle ambition among women who, in a less sexist media environment, would be willing to stand as political candidates.     

A CANONICAL ANALYSIS OF MANAGERIAL-ENVIRONMENTAL RELATIONSHIPS: IMPLICATIONS FOR METHODOLOGY AND POLICY*

The present study uses canonical analysis to examine the contingency model of organizations in the nursing home field. Data on environmental pressures and managerial practices relevant to the decision-making process are analyzed. Results support the theory that managerial practices are contingent on environmental pressures. The implications of these results for health-care decision makers and for future assessments of managerial-environmental relationships are discussed.     

Issuing and Responding to Unusual Questions: A Conversation Analytic Account of Tom Andersen's Therapeutic Practice

Tom Andersen is considered one of the key contributors to the development of postmodern practice. Little is known, however, about the ways in which his ideas and practices are routinely carried out in situ. We used Conversation Analysis (CA) to investigate a session of couple therapy facilitated by Andersen. We show how Andersen and client participants oriented to and addressed problems of understanding that occurred between them. The source of this trouble was Andersen's use of unusual question formulations. We offer preliminary evidence that such unusual formulations served particular interactional and institutional (i.e., therapeutic) functions in their local contexts of use. We conclude by considering some implications of this analysis—and of conversation analytic inquiry more generally—for the practice of family therapy.