Illusions in regression analysis

Armstrong, J. Scott

doi:10.1016/j.ijforecast.2012.02.001

Cited by 275 publications

(129 citation statements)

References 33 publications

Supporting

Mentioning

117

Contrasting

Unclassified

Order By: Relevance

“…Armstrong (2012) pointed out that achieving high fit validity with a regression model is easy since such models use all available information in the data to build the model to test the model; he demonstrated the point by estimating a regression model using data from a table of random numbers that achieves high statistical significance.…”

Section: Testing For Predictive Validity With Additional (Holdout) Samentioning

confidence: 99%

Constructing Business-to-Business Marketing Models that Overcome the Limitations in Variable-Based and Case-Based Research Paradigms

Woodside

2015

Journal of Business-to-Business Marketing

View full text Add to dashboard Cite

Purpose: This article describes the principal limitations frequently observable in variable-based and case-based research in business-to-business (B-to-B) marketing.Focus: The limitations relate to B-to-B theory construction, data collection, and data analysis and interpretation. The principal limitations in variable-based research include the reliance on regression (net effects) models in theory construction and the pervasive failure to test resulting empirical regression models for predictive validity. The principal limitations of case-based research are the frequent absence of rigorous theory formulation either pre-or postdata collection and failures to test propositions for predictive validity.Recommendations: The article directs attention to iconic studies that break free in part from these and additional limitations. The article offers useful steps on how to break free completely from the principal limitations.Managerial implications: B-to-B executives should insist that researchers include asymmetric algorithm models in reports that include symmetric (regression analysis) and analysis of variance findings.

show abstract

Section: Testing For Predictive Validity With Additional (Holdout) Samentioning

confidence: 99%

Constructing Business-to-Business Marketing Models that Overcome the Limitations in Variable-Based and Case-Based Research Paradigms

Woodside

2015

Journal of Business-to-Business Marketing

View full text Add to dashboard Cite

show abstract

“…High fit validities for models built using isomorphic or regression analysis tools are easy to achieve, and inadequate. Armstrong (2012) and Gigerenzer and Brighton (2009) provide convincing evidence that fit-validityonly tests are inadequate and frequently indicate inaccurate predictions of the behaviors/ decisions of cases in hold-out samples. Howard and Morgenroth (1968) provide asymmetric decision recipes for high price and low price decisions.…”

Section: Journal Of Global Scholars Of Marketing Science 261mentioning

confidence: 97%

“…The illusions of regression analysis because of over-fitting of models are points emphasized by Armstrong (1985Armstrong ( , 2012) that most researchers continue to ignore. Gigerenzer and Brighton (2009, p. 118) go beyond pointing out the severe limitations of regression A.G. Woodside 262 analysis following their statement that "achieving a good fit to observations does not necessarily mean we have found a good model, and choosing the model with the best fit is likely to result in poor predictions.…”

Section: Statistical Power and Effect Size In Marketing Researchmentioning

confidence: 99%

Iconic studies relevant for research in marketing and theJournal of Global Scholars of Marketing Science

Woodside

2015

Journal of Global Scholars of Marketing Science

View full text Add to dashboard Cite

Here, an "icon" refers to a study widely and critically admired, a study symbolizing a movement or field of activity; iconic marketing studies offer exceptional contributions to marketing theory and/or data collection, and/or data analysis, and/or interpretations/ implications of qualitative and/or quantitative empirical findings. This article identifies antecedent conditions associated with achieving iconic status in research in marketing and proposes 10 tenets useful for identifying and planning iconic studies in marketing. The study reviews examples of iconic studies in the marketing literature. The study also addresses a few telling mistakes that researchers in the marketing discipline frequently make. The study of iconic research is helpful for crafting high-quality theory and planning high-quality research designs, as well as increasing vigilance and skill in identifying truly exceptionally high-quality studies and studies that are plainly just bad. The essay briefly reviews 10 JGSMS articles appearing also in this virtual issue as possible candidates for the achievement of iconic status in marketing.

show abstract

“…The purpose of its analysis is to predict the output variables with the value of multiple explanatory variables. The main limitation of the model is that the correlation between the variables changes with time and space [48]. Assuming an output variable is y i , and some explanatory variables are x i , then the relationship between the output variable and the explanatory variable can be expressed as:…”

Section: Multiple Linear Regression Modelmentioning

confidence: 99%

Forecasting of Chinese Primary Energy Consumption in 2021 with GRU Artificial Neural Network

Liu

Bielefield

et al. 2017

Energies

View full text Add to dashboard Cite

Abstract:The forecasting of energy consumption in China is a key requirement for achieving national energy security and energy planning. In this study, multi-variable linear regression (MLR) and support vector regression (SVR) were utilized with a gated recurrent unit (GRU) artificial neural network of Chinese energy to establish a forecasting model. The derived model was validated through four economic variables; the gross domestic product (GDP), population, imports, and exports. The performance of various forecasting models was assessed via MAPE and RMSE, and three scenarios were configured based on different sources of variable data. In predicting Chinese energy consumption from 2015 to 2021, results from the established GRU model of the highest predictive accuracy showed that Chinese energy consumption would be likely to fluctuate from 2954.04 Mtoe to 5618.67 Mtoe in 2021.

show abstract

Illusions in regression analysis

Cited by 275 publications

References 33 publications

Constructing Business-to-Business Marketing Models that Overcome the Limitations in Variable-Based and Case-Based Research Paradigms

Constructing Business-to-Business Marketing Models that Overcome the Limitations in Variable-Based and Case-Based Research Paradigms

Iconic studies relevant for research in marketing and theJournal of Global Scholars of Marketing Science

Forecasting of Chinese Primary Energy Consumption in 2021 with GRU Artificial Neural Network

Contact Info

Product

Resources

About