<b>Research Commentary</b>—Too Big to Fail: Large Samples and the <i>p</i>-Value Problem

Lin, Ming-Chyuan; Lucas, Henry C.; Shmueli, Galit

doi:10.1287/isre.2013.0480

Cited by 810 publications

(562 citation statements)

References 49 publications

Supporting

Mentioning

541

Contrasting

Unclassified

Order By: Relevance

“…Therefore, based on recent work on this topic Lin et al 2013, in Table 5, we present the effect size and 95% confidence interval for both direct effects and moderation effects. Specifically, one unit increase in RM, WIP, and FG inventory, ROA increases by 0.86, 0.48, and 0.43 percent, respectively.…”

Section: Analysis and Resultsmentioning

confidence: 99%

Interdependence Among Inventory Types and Firm Performance

Manikas

2017

OSCM: An Int. Journal

View full text Add to dashboard Cite

Extant research on lean manufacturing has focused on the relationship between systemic inventory and performance. However, limited research, if any, has focused on dynamics between three main types of inventory -raw material, workin-process, and finished goods -and firm performance. Drawing on the interdependence framework and a sample of 1,286 firms representing 41,067 firm-quarter observations from 2000 to 2013, when considering two types of inventory efficiencies jointly, higher efficiency in the subsequent inventory type must be accompanied by lower efficiency in the previous inventory type. In other words, lower raw material inventory efficiency and higher work-in-process inventory efficiency, or lower work-in-process inventory efficiency and higher finished goods inventory efficiency are associated with higher performance. High performance is also realized only when inventory efficiency is increasing across all three inventory types. The findings show that pursuing efficiency in one type of inventory requires lower efficiency in the previous inventory stage, and benefits of lean manufacturing could only be realized when all three inventory efficiencies are simultaneously high.

show abstract

Section: Analysis and Resultsmentioning

confidence: 99%

Interdependence Among Inventory Types and Firm Performance

Manikas

2017

OSCM: An Int. Journal

View full text Add to dashboard Cite

show abstract

“…The low p-values are counter-intuitive given the low D-statistics. We observe from the last columns of Table 5 and Table 6 that these distributions have large sample size, suggesting that the low p-values are an artifact of a large number of samples [15]. To support this claim, we compare the EDFs with their respective input CDFs.…”

Section: Validationmentioning

confidence: 88%

REGENT: A Framework for Realistic Generation of Network Traffic

Divakaran¹,

Fok²,

Ko³

et al. 2016

IJCCE

View full text Add to dashboard Cite

Network security research is highly dependent on traffic datasets. Fair and comprehensive analysis as well as performance evaluations of different solutions for problems such as detections of intrusions, anomalies and attacks, requires traffic traces that closely resemble data from operational network. Currently available datasets are either getting obsolete with time, or lacking important information such as ground truth of data and payloads in the traffic. While making little compromises, an alternate solution to this problem is to generate traffic data. However, care has to be taken that such a solution is capable to cope up with the changing characteristics of traffic; more generally, it should be flexible enough to generate traffic with specific characteristics as required by a user. In this work, we develop a framework for realistic generation of network traffic, called REGENT, which takes traffic models as input. In REGENT, different protocols generate real traffic independently, and based on the specific models (such as distribution for inter-arrival time between connections, distribution for connection size, etc.) provided by a user. We conduct experiments wherein REGENT takes protocol models as input, and generates real traffic as output. Using analysis, we show that the characteristics of the generated traffic (protocols) are close to the models specified as input.

show abstract

“…Due to the large sample size, p values were not used for between-model comparisons, as even negligible differences are likely to be statistically significant. 34 Instead, as adjunctive measures of model performance, the Akaike information criterion (AIC), Bayesian information criterion (BIC), pseudo R 2 , and shrinkage coefficients were also calculated. Smaller AIC and BIC values and larger pseudo R 2 and shrinkage coefficients indicate better model performance.…”

Section: Discussionmentioning

confidence: 99%

Functional Status Outperforms Comorbidities in Predicting Acute Care Readmissions in Medically Complex Patients

Shih

Gerrard²,

Goldstein

et al. 2015

J GEN INTERN MED

View full text Add to dashboard Cite

OBJECTIVE:To examine functional status versus medical comorbidities as predictors of acute care readmissions in medically complex patients. DESIGN: Retrospective database study. SETTING: U.S. inpatient rehabilitation facilities. PARTICIPANTS: Subjects included 120,957 patients in the Uniform Data System for Medical Rehabilitation admitted to inpatient rehabilitation facilities under the medically complex impairment group code between 2002 and 2011. INTERVENTIONS: A Basic Model based on gender and functional status was developed using logistic regression to predict the odds of 3-, 7-, and 30-day readmission from inpatient rehabilitation facilities to acute care hospitals. Functional status was measured by the FIM ® motor score. The Basic Model was compared to six other predictive models-three Basic Plus Models that added a comorbidity measure to the Basic Model and three GenderComorbidity Models that included only gender and a comorbidity measure. The three comorbidity measures used were the Elixhauser index, Deyo-Charlson index, and Medicare comorbidity tier system. The c-statistic was the primary measure of model performance. MAIN OUTCOME MEASURES: We investigated 3-, 7-, and 30-day readmission to acute care hospitals from inpatient rehabilitation facilities. RESULTS: Basic Model c-statistics predicting 3-, 7-, and 30-day readmissions were 0.69, 0.64, and 0.65, respectively. The best-performing Basic Plus Model (Basic+Elixhauser) c-statistics were only 0.02 better than the Basic Model, and the best-performing GenderComorbidity Model (Gender+Elixhauser) c-statistics were more than 0.07 worse than the Basic Model. CONCLUSIONS:Readmission models based on functional status consistently outperform models based on medical comorbidities. There is opportunity to improve current national readmission risk models to more accurately predict readmissions by incorporating functional data.

show abstract

Research Commentary—Too Big to Fail: Large Samples and the p-Value Problem

Cited by 810 publications

References 49 publications

Interdependence Among Inventory Types and Firm Performance

Interdependence Among Inventory Types and Firm Performance

REGENT: A Framework for Realistic Generation of Network Traffic

Functional Status Outperforms Comorbidities in Predicting Acute Care Readmissions in Medically Complex Patients

Contact Info

Product

Resources

About