Consistency: a pre-requisite for inspecting old people's homes?

Correspondence to Dr Ian Gibbs, Centre for Housing Policy, University of York, Heslington, York YO1 5DD. Summary Inspection of old people's homes, in both local authority andindependent sectors, forms a major component in the work ofthe new ‘arm's length‘ inspection units. The successof these units will depend in part on inspectors' ability tomake consistent judgements based on agreed standards. This articlereports on a study which examined the extent to which inspectorswere able to reach reliable conclusions about the quality ofcare in old people's homes and the possible contribution ofchecklists in this task.     

MEASURING LEVELS OF PARTY IDENTIFICATION DOES QUESTION ORDER MATTER?

This article examines whether asking the vote question beforeparty identification alters the strength of partisanship andits relationship to vote choice. It employs the 1992 BritishElection Survey, which included a random split half-sample experiment,and the 1992–93 American Election Study Panel, where thequestion order for party identification and the vote were changed.The results show that altering the question ordering had verylittle effect in Britain and no significant effect in the UnitedStates. These results are consistent with the notion that partyidentification is one of the more enduring and stable componentsof mass political behavior in both presidential and parliamentarysystems.     

Windows 2000 Server典型实验在VMware中的实现

Windows 2000 Server的实验需要专门的网络实验室,从而需要很大的设备投入。随着虚拟仿真技术在计算机网络中的应用,可以利用虚拟机软件构造虚拟计算机网络实验平台。VMware是一个功能强大的虚拟机软件,本文介绍VMware环境下Windows2000 Server的多个典型实验的实现原理与方法。     

Selecting for Social Work

Summary Social work courses need to spell out the characteristics andconsequences of particular selection procedures, as a prerequisiteof improvement. This article reports a survey of applicantsfor a postgraduate social work course, where entry is determinedby two independent, individual interviews. Previous social workexperience and the candidate's degree subject emerged as importantcorreiates of selection. Paired interviewing appeared to operatein an impartial fashion, with a high degree of consistency betweeninterviewers of different kinds, and an absence of hierarchyin resolving conflicts. However, the outcome of interviewingwas partially contingent upon the degree of agreement betweeninterviewers on their initial, individual assessment of thecandidates' suitability.     

Agency Decentralization: Planning and Assessment

Summary Criteria are examined for the location of area offices in SocialServices Departments, drawing on the monitored experience ofdealing with voluntary callers to an area office in the ProbationService. Changes were observed in numbers of voluntary callersand the kinds of problems presented, together with developmentsin inter-agency relationships. The caller's subjective feelingsregarding the accessibility of the agency were found to be importantdeterminants of agency use. Implications are drawn out for planningdecisions in local authorities     

Campaign Activities and Electoral Outcomes in Britain 1979 and 1983

Political broadcasting through the electronic media has largelydisplaced two older methods of influencing voters—canvassingand advertising in the press—as the dominant party campaignactivity during elections. This article utilizes survey datacollected during the 1979 and 1983 British general electionsto estimate the reported effects of these three types of campaignactivity on electoral outcomes. The results show that few votersreport being influenced by canvassing and advertising, and thenet effect on the vote is either small or nonexistent. By contrast,around 1 in 6 report being influenced by a party political broadcast,and those most influenced are eventual Liberal-Alliance voters.The net electoral effect of political broadcasting can rangeas high as 3.4 percent, but gains for the Conservatives andLabour are offset by the political broadcasting of competitors.The main exceptions are the Liberals in 1979 and the Alliancein 1983, who made a net gain of 1.5 percent and 3.3 percentof the vote, respectively.     

Statistical inference concerning weighted Poisson rates under some natural order restrictions

A conditional Poisson model is proposed for situations where one wants to compare the rate parameters for several populations, adjusting for a 'size' parameter and the random e ects of the subjects. Sometimes the physical nature of the problem makes it logical to consider some order restrictions on the rate parameters. The rate parameters are estimated and tested for under such order restrictions. This investigation is motivated by a real-life example on butterfly behavior, and the estimation and tests of hypotheses are illustrated over this data set.