Search results

To explore the appropriateness of statistical significance testing to measure the practical, managerial significance of outcomes in marketing programmes.

An in‐depth analysis of SST's scientific roots is coupled with delineation of a set of general objectives of marketing‐programme measurement to identify the applicability limits of significance testing.

In particular, it is shown that the relatively well known sample‐size dependence of SST and its somewhat lesser known replicability, representativeness and impact fallacies can severely affect the robustness of significance tests. Statistical significance is not the same concept as practical significance.

Comprehensive discussion of principles and practice leads to a set of prescriptive usage recommendations, directed at the goal of establishing much‐needed applicability rules and limits for the use of significance‐testing methodologies in an applied marketing context.

This robust challenge to the efficacy of significance testing in marketing practice should be of interest to any marketing planner concerned with the collection and use of marketing intelligence.

Organizational decision‐makers require information presented in ways that allow them to make informed decisions on the effectiveness of change interventions. Current statistical methods do not provide enough information about the practical value of organizational interventions to decision‐makers. It is proposed that a strong hypothesis testing strategy provides a partial answer to this problem. The hypothesis testing method presented here uses Bayesian statistics to test null hypotheses other than the traditional Ho = 0. A description of the evaluation of a change project in six manufacturing plants of a large United States corporation is provided. The data from this project is used to show how both statistical and practical significance may be tested using this hypothesis testing method. The applicability of the strong hypothesis testing approach to the assessment of organizational change is then discussed, and recommendations are made for evaluations conducted in field settings.

Discusses alternative strategies which may be employed when differences arise between achieved audit‐sampling results and planned results, which means that risk levels used in ex post decision making may be different from planned levels. Contrasts a conventional strategy — which is to fix the risk of incorrect acceptance at a planned level and to ignore the risk of incorrect rejection or to accept the minimum available level of that risk which is consistent, after the fact, with the planned level of risk of incorrect acceptance — with a theoretically appealing strategy which balances both risk levels in proportion to their perceived disutility. Reports on the results of an experiment involving these two strategies, in which all subjects were auditors with statistical audit experience. Suggests that the most important statistically significant finding is that, in certain circumstances, these auditors are more willing to base audit decisions on statistical evidence after the alternative strategy is explained and available for their use.

Describes statistical methods applied to sensory discrimination tests. Illustrates binomial and chi‐square statistical analysis and discusses similarity testing, power and replication in discrimination testing.

The purpose of this paper is to consider the concepts of individual and complete statistical power used for multiple testing and shows their relevance for determining the number of statistical tests to perform when assessing non-response bias.

A statistical power analysis of 55 survey-based research papers published in three prestigious logistics journals (International Journal of Physical Distribution and Logistics Management, Journal of Business Logistics, Transportation Journal) over the last decade was conducted.

Results show that some of the low complete power levels encountered could have been avoided if fewer tests had been used in the assessment of non-response bias.

The research offers important recommendations to scholars engaged in survey research as they assess the effects of non-respondents on research findings. By following the recommended strategies for testing non-response bias, researchers can improve the statistical power of their findings.

The overall goal of science is to build a valid and reliable body of knowledge about the functioning of the world and how applying that knowledge can change it. As personnel and human resources management researchers, we aim to contribute to the respective bodies of knowledge to provide both employers and employees with a workable foundation to help with those problems they are confronted with. However, what research on research has consistently demonstrated is that the scientific endeavor possesses existential issues including a substantial lack of (a) solid theory, (b) replicability, (c) reproducibility, (d) proper and generalizable samples, (e) sufficient quality control (i.e., peer review), (f) robust and trustworthy statistical results, (g) availability of research, and (h) sufficient practical implications. In this chapter, we first sing a song of sorrow regarding the current state of the social sciences in general and personnel and human resources management specifically. Then, we investigate potential grievances that might have led to it (i.e., questionable research practices, misplaced incentives), only to end with a verse of hope by outlining an avenue for betterment (i.e., open science and policy changes at multiple levels).

Generator maintenance scheduling (GMS) is an essential task for electric power utilities as the periodical maintenance activity enhances the lifetime and also ensures the reliable and continuous operation of generating units. Though numerous meta-heuristic algorithms have been reported for the GMS solution, enhancing the existing techniques or developing new optimization procedure is still an interesting research task. The meta-heuristic algorithms are population based and the selection of their algorithmic parameters influences the quality of the solution. This paper aims to propose statistical tests guided meta-heuristic algorithm for solving the GMS problems.

The intricacy characteristics of the GMS problem in power systems necessitate an efficient and robust optimization tool. Though several meta-heuristic algorithms have been applied to solve the chosen power system operational problem, tuning of their control parameters is a protracting process. To prevail over the previously mentioned drawback, the modern meta-heuristic algorithm, namely, ant lion optimizer (ALO), is chosen as the optimization tool for solving the GMS problem.

The meta-heuristic algorithms are population based and require proper selection of algorithmic parameters. In this work, the ANOVA (analysis of variance) tool is proposed for selecting the most feasible decisive parameters in algorithm domain, and the statistical tests-based validation of solution quality is described. The parametric and non-parametric statistical tests are also performed to validate the selection of ALO against the various competing algorithms. The numerical and statistical results confirm that ALO is a promising tool for solving the GMS problems.

As a first attempt, ALO is applied to solve the GMS problem. Moreover, the ANOVA-based parameter selection is proposed and the statistical tests such as Wilcoxon signed rank and one-way ANOVA are conducted to validate the applicability of the intended optimization tool. The contribution of the paper can be summarized in two folds: the ANOVA-based ALO for GMS applications and statistical tests-based performance evaluation of intended algorithm.

Attaining high response rates in survey‐based supply chain management (SCM) research is becoming increasingly difficult, but small samples can limit the reliability and validity of empirical research findings. The purpose of this article is to analyze the status quo and provide a discussion of methodological issues related to the use of small samples in SCM research.

An in‐depth review of 75 small sample survey studies published between 1998 and 2007 in three journals in the field that frequently publish survey‐based research papers (TJ, IJPDLM, and JBL) was conducted, and key characteristics of these studies were compared with the characteristics from 44 small sample survey studies published in leading operations management (JOM) and management (AMJ) journals.

The review of papers published in TJ, IJPDLM, and JBL shows that small samples are frequently used in SCM research. This study provides an overview of current practices, opportunities for improvement, and a number of specific recommendations that may help increase the analytical rigor of (future) survey‐based studies that rely on small samples.

The recommendations provided in this article can greatly benefit researchers in the field of SCM. By following these proposals, the reliability and validity of research findings will be increased, researchers will be better equipped to investigate interesting questions where small samples are the norm rather than the exception (e.g., the study of dyadic supply chain relationships), and important and valid contributions to the theory and practice of SCM will be generated.