Comparing the overlapping of two independent confidence intervals with a single confidence interval for two normal population parameters

Maghsoodloo, Saeed; Huang, Ching Ying

doi:10.1016/j.jspi.2010.04.057

Cited by 36 publications

(27 citation statements)

References 4 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…differences between groups is required (results of such a test are provided in Table A.2). See Maghsoodloo and Huang (2010) for a discussion of this issue and the related literature.…”

Section: Differences Between Institutionsmentioning

confidence: 99%

Do UK Higher Education Students Overestimate Their Starting Salary?*

Jerrim¹

2011

Fiscal Studies

View full text Add to dashboard Cite

The wage expectations of university students have relevance for human capital theory, models of student enrolment, and public policy on the provision of higher education. However, these expectations have been the subject of relatively little research in European countries, with no contemporary evidence available in the UK. There has also been little consideration in the literature thus far of how much estimates may be biased by non‐response and sample selection. This paper thus makes an original contribution to the literature by comparing UK undergraduate students' wage expectations with the actual wages earned by the same cohort on graduation, after attempting to correct for the aforementioned statistical issues in terms of observable characteristics. In contrast to other European studies, I conclude that, on average, full‐time students overestimate their starting salary. Yet I also find substantial variation in the extent of overestimation between different subgroups.

show abstract

“…differences between groups is required (results of such a test are provided in Table A.2). See Maghsoodloo and Huang (2010) for a discussion of this issue and the related literature.…”

Section: Differences Between Institutionsmentioning

confidence: 99%

Do UK Higher Education Students Overestimate Their Starting Salary?*

Jerrim¹

2011

Fiscal Studies

View full text Add to dashboard Cite

show abstract

“…It has been shown elsewhere that when making statistical inferences from the comparison of confidence (credible) intervals, using 95% intervals is an overly conservative test for statistical differences (Goldstein & Healy, 1995). Rather, attaining a conventional type I error rate of 5% requires the use of approximately 83.5% intervals (Maghsoodloo & Huang, 2010; Payton, Greenstone, & Schenker, 2003). 14 We expect to find no statistically discernible difference in the predicted evaluations for the high‐NFC/low‐NFA type across primes (i.e., a prediction of overlap in the credible intervals).…”

Section: Methodsmentioning

confidence: 99%

The Effects of Need for Cognition and Need for Affect on Partisan Evaluations

Arceneaux

Wielen

2012

Political Psychology

View full text Add to dashboard Cite

The voting behavior literature has advanced two prominent theoretical models of partisanship: the social psychological and rational models. Implicit to both stylized models is the assumption that all partisans process information similarly. Yet, growing research in psychology suggests that individuals possess different motivations when evaluating information. We propose that the applicability of the stylized models of partisanship is conditioned on individuals' need for cognition (NFC) and need for affect (NFA), with the social psychological model being most applicable to individuals who have a high NFA and the rational model most applicable to those with a high NFC. To test this proposition, we fielded a survey in which respondents who identified with the two major political parties in the United States (Democrat or Republican) were randomly assigned factual information that depicted either their party or their opposing party in a negative light. Respondents were then asked to assess the actions of that party and subsequently evaluate both political parties. We find evidence that is generally consistent with the proposition that the stylized models of partisanship are conditionally dependent on the extent to which individuals possess a need to engage in effortful thinking or a need to seek out emotions.

show abstract

“…Certainly simply comparing model estimates, such as among IC50s (dose creating a 50% inhibition of the reaction) and their standard errors or 95% confidence limits for overlap is an unacceptable approach. For a discussion of the relevant issues, refer to Payton and others (Maghsoodloo and Huang, 2010;Payton et al, 2003;Schenker and Gentleman, 2001). For the practicing biomedical scientist, even their exploratory methods have significant limitations including that they require highly technical code to calculate, and are not yet generally accepted.…”

Section: Generalmentioning

confidence: 99%

Microplate fluorescence protease assays test the inhibition of select North American snake venoms' activities with an anti-proteinase library

Price

2015

Toxicon

View full text Add to dashboard Cite

Comparing the overlapping of two independent confidence intervals with a single confidence interval for two normal population parameters

Cited by 36 publications

References 4 publications

Do UK Higher Education Students Overestimate Their Starting Salary?*

Do UK Higher Education Students Overestimate Their Starting Salary?*

The Effects of Need for Cognition and Need for Affect on Partisan Evaluations

Microplate fluorescence protease assays test the inhibition of select North American snake venoms' activities with an anti-proteinase library

Contact Info

Product

Resources

About