Sample Size Determination for Logistic Regression: A Simulation Study

Influence of student's ability to delay gratification on their educational transition choice

Objective

A considerable amount of research identified socio-economic status and cognitive ability as robust predictors, the influence of student's ability to delay gratification (ADG) on their educational transition choice doesn't received researcher's attention. To address this gap, the present study examined the incremental power of students ADG in predicting the dichotomous choice i.e. the choice of general or vocational education after successful completion of compulsory schooling.

Methods

Amid Covid-19 pandemic, cross sectional survey via an online mode was found feasible for the data collection process in our study. An online link of survey questionnaire was created in the Google forms and administered to (N = 1024) grade 8 students in the Union Territory of Jammu & Kashmir, India. Multiple binary logistic regressions were conducted to predict the students' choice, and odds ratios and average marginal effects were reported for better interpretation of results.

Results

Our results showed that students tracking choice differed significantly with respect to their gender and locale (smaller effect), ADG (medium effect), and cognitive ability and socio-economic status (larger effect). The probability of choosing the track of vocational education (with general education track as a baseline category) increases as students ADG decreases, and vice versa. This association of student's ADG with the choice of vocational education track held same over and above the covariates-socio-economic status, cognitive ability, gender and locale.

Supplementary Information

The online version contains supplementary material available at 10.1186/s40461-022-00134-6.

Energy Justice in Slum Rehabilitation Housing: An Empirical Exploration of Built Environment Effects on Socio-Cultural Energy Demand

The interaction of energy and buildings institutes a complex socio-technical system that influences the eudemonic well-being of the occupants. Understanding these drivers become even more necessary in impoverished areas where occupants struggle to avail essential energy services. The literature indicates that energy injustice can be addressed through provisioning of comfort, cleanliness, and convenience (3Cs) as critical cultural energy services in low-income areas. This study investigates the socio-architectural influence for slum rehabilitation housing (SRH) on cultural energy services that can promote distributive justice. The methodology adopts an empirical route using data from 200 household surveys from SRH in Mumbai, India, and João Pessoa, Brazil. A model between the 3Cs and socio-architectural elements was established using Firth’s binary logistic regression. The survey results showed that the SRH in Brazil had twice the appliance ownership as compared to the Mumbai SRH. There were distinct energy service preferences in the study areas, despite common poverty burdens. The empirical results showed that the lack of socio-architectural design elements like open spaces, privacy, and walkability in the study areas demanded specific comfort and convenience appliances as a counter-response. A critical policy implication drawn was on the need for socio-architectural inclusive energy planning for distributive justice in poverty. Mitigating rising energy demand through appropriate built environment design of slum rehabilitation housing can contribute to fulfilling the UN’s SDG 7 (clean and affordable energy) and 11 (sustainable cities and communities) goals.

Simulation-based sample-sizing and power calculations in logistic regression with partial prior information

There have been many approximations developed for sample sizing of a logistic regression model with a single normally-distributed stimulus. Despite this, it has been recognised that there is no consensus as to the best method. In pharmaceutical drug development, simulation provides a powerful tool to characterise the operating characteristics of complex adaptive designs and is an ideal method for determining the sample size for such a problem. In this paper, we address some issues associated with applying simulation to determine the sample size for a given power in the context of logistic regression. These include efficient methods for evaluating the convolution of a logistic function and a normal density and an efficient heuristic approach to searching for the appropriate sample size. We illustrate our approach with three case studies. Copyright © 2016 John Wiley & Sons, Ltd.