Do Larger Sample Sizes Increase the Reliability of Traffic Incident Duration Models? A Case Study of East Tennessee Incidents

2021 ◽  
pp. 036119812199206
Author(s):  
Zihe Zhang ◽  
Jun Liu ◽  
Xiaobing Li ◽  
Asad J. Khattak
Keyword(s):  
Data Collection ◽  
Sample Size ◽  
Information Dissemination ◽  
Threshold Value ◽  
Sample Size Determination ◽  
Duration Models ◽  
Critical Threshold ◽  
Sample Sizes ◽  
Incident Duration

Incident duration models are often developed to assist incident management and traveler information dissemination. With recent advances in data collection and management, enormous achieved incident data are now available for incident model development. However, a large volume of data may present challenges to practitioners, such as data processing and computation. Besides, data that span multiple years may have inconsistency issues because of the data collection environments and procedures. A practical question may arise in the incident modeling community—Is that much data really necessary (“all-in”) to build models? If not, then how many data are necessary? To answer these questions, this study aims to investigate the relationship between the data sample sizes and the reliability of incident duration analysis models. This study proposed and demonstrated a sample size determination framework through a case study using data of over 47,000 incidents. This study estimated handfuls of hazard-based duration models with varying sample sizes. The relationships between sample size and model performance, along with estimate outcomes (i.e., coefficients and significance levels), were examined and visualized. The results showed that the variation of estimated coefficients decreases as the sample size increases, and becomes stabilized when the sample size reaches a critical threshold value. This critical threshold value may be the recommended sample size. The case study suggested a sample size of 6,500 to be enough for a reliable incident duration model. The critical value may vary significantly with different data and model specifications. More implications are discussed in the paper.

scholarly journals Adequacy of sample size in a qualitative case study and the dilemma of data saturation: A narrative review

2021 ◽  
Vol 10 (3) ◽  
pp. 180-187
Author(s):  
Felix Chukwuma Aguboshim
Keyword(s):  
Data Collection ◽  
Sample Size ◽  
Qualitative Case Study ◽  
Study Data ◽  
Narrative Review ◽  
Sample Sizes ◽  
Significant Information ◽  
Probability Sample ◽  
Data Saturation

The consensus of many researchers on data saturation is that data saturation is a key driver for determining the adequacy of sample size in a qualitative case study. Despite these global consensuses, some researchers described data saturation as complex because the decision to stop data collection is solely dictated by the judgment and experience of researchers. Other researchers claimed that guidelines for determining non-probability sample sizes, used as an indication of data saturation are virtually non-existent, problematic, or controversial. Others claimed that data saturation hitched to sample size is practically weak, because data are never truly saturated, as there are always new data to be discovered. This narrative study highlights the dilemma of data saturation and strategies to adequately determine sample size in a qualitative case study. A narrative review of prior research that focused on the vast works of literature that revealed significant information on data saturation and strategies to adequately determine sample size was adopted. Peer-reviewed articles within the last five years from electronic databases, using some keywords such as “qualitative case study”, “sample size in a qualitative case study”, “data saturation”, etc., were also extracted. Results show that data saturation is very helpful especially at the conceptual stage, but its concept and standard is elusive, because it lacks practical guidance for estimating sample size for a robust research prior to data collection. Findings from this study may encourage researcher on better guidelines for determining non-probability sample sizes.

scholarly journals Sample Size Determination and Optimal Design of Simple Pretest-Posttest Experimental Designs: Introduction, Software, and Illustrations

2021 ◽  
Author(s):  
Metin Bulus
Keyword(s):  
Optimal Design ◽  
Sample Size ◽  
Small Sample ◽  
Experimental Designs ◽  
Sample Size Determination ◽  
Size Determination ◽  
Sample Sizes ◽  
Control Groups ◽  
Small Sample Sizes ◽  
And Control

A recent systematic review of experimental studies conducted in Turkey between 2010 and 2020 reported that small sample sizes had been a significant drawback (Bulus and Koyuncu, 2021). A small chunk of the studies were small-scale true experiments (subjects randomized into the treatment and control groups). The remaining studies consisted of quasi-experiments (subjects in treatment and control groups were matched on pretest or other covariates) and weak experiments (neither randomized nor matched but had the control group). They had an average sample size below 70 for different domains and outcomes. These small sample sizes imply a strong (and perhaps erroneous) assumption about the minimum relevant effect size (MRES) of intervention before an experiment is conducted; that is, a standardized intervention effect of Cohen’s d < 0.50 is not relevant to education policy or practice. Thus, an introduction to sample size determination for pretest-posttest simple experimental designs is warranted. This study describes nuts and bolts of sample size determination, derives expressions for optimal design under differential cost per treatment and control units, provide convenient tables to guide sample size decisions for MRES values between 0.20 ≤ Cohen’s d ≤ 0.50, and describe the relevant software along with illustrations.

scholarly journals Adaptive sample size determination for the development of clinical prediction models

2020 ◽  
Author(s):  
Evangelia Christodoulou ◽  
Maarten van Smeden ◽  
Michael Edlinger ◽  
Dirk Timmerman ◽  
Maria Wanitschek ◽  
...  
Keyword(s):  
Ovarian Cancer ◽  
Sample Size ◽  
Prediction Models ◽  
A Priori ◽  
Sample Size Calculation ◽  
Model Performance ◽  
Sample Size Determination ◽  
Sample Sizes ◽  
Cancer Data ◽  
C Statistic

Abstract Background: We suggest an adaptive sample size calculation method for developing clinical prediction models, in which model performance is monitored sequentially as new data comes in. Methods: We illustrate the approach using data for the diagnosis of ovarian cancer (n=5914, 33% event fraction) and obstructive coronary artery disease (CAD; n=4888, 44% event fraction). We used logistic regression to develop a prediction model consisting only of a-priori selected predictors and assumed linear relations for continuous predictors. We mimicked prospective patient recruitment by developing the model on 100 randomly selected patients, and we used bootstrapping to internally validate the model. We sequentially added 50 random new patients until we reached a sample size of 3000, and re-estimated model performance at each step. We examined the required sample size for satisfying the following stopping rule: obtaining a calibration slope ≥0.9 and optimism in the c-statistic (ΔAUC) <=0.02 at two consecutive sample sizes. This procedure was repeated 500 times. We also investigated the impact of alternative modeling strategies: modeling nonlinear relations for continuous predictors, and applying Firth’s bias correction.Results: Better discrimination was achieved in the ovarian cancer data (c-statistic 0.9 with 7 predictors) than in the CAD data (c-statistic 0.7 with 11 predictors). Adequate calibration and limited optimism in discrimination was achieved after a median of 450 patients (interquartile range 450-500) for the ovarian cancer data (22 events per parameter (EPP), 20-24), and 750 patients (700-800) for the CAD data (30 EPP, 28-33). A stricter criterion, requiring ΔAUC <=0.01, was met with a median of 500 (23 EPP) and 1350 (54 EPP) patients, respectively. These sample sizes were much higher than the well-known 10 EPP rule of thumb and slightly higher than a recently published fixed sample size calculation method by Riley et al. Higher sample sizes were required when nonlinear relationships were modeled, and lower sample sizes when Firth’s correction was used. Conclusions: Adaptive sample size determination can be a useful supplement to a priori sample size calculations, because it allows to further tailor the sample size to the specific prediction modeling context in a dynamic fashion.

Reconsideration of Sample Size Requirements for Field Traffic Data Collection with Global Positioning System Devices

2002 ◽  
Vol 1804 (1) ◽  
pp. 17-22 ◽  
Author(s):  
Shuo Li ◽  
Karen Zhu ◽  
B. H. W. van Gelder ◽  
John Nagle ◽  
Carl Tuttle
Keyword(s):  
Data Collection ◽  
Sample Size ◽  
Travel Time ◽  
Global Positioning System ◽  
Work Zone ◽  
Traffic Data ◽  
Sample Sizes ◽  
Traffic Data Collection ◽  
Global Positioning ◽  
Gps Devices

The use of Global Positioning System (GPS) technologies has expanded to perform traffic data collection for transportation studies such as work zone studies. To generate reliable results from the data acquired by using GPS devices, it is necessary to investigate such factors as sample size requirements that may affect a specific study and to establish a consistent method for data collection. It has been confirmed that the Institute of Transportation Engineers’ Manual of Transportation Engineering Studies usually underestimates the sample sizes for travel time and delay studies. However, the hybrid method developed by Quiroga and Darcy overestimates the sample sizes. A modified equation is presented to estimate the minimum sample sizes for collecting field data with GPS devices. Travel speed may be more stable and can be easily measured for travel time and delay studies. Stopped delay varies considerably at intersections, and the sample sizes depend to a large extent on the permitted errors. Work zone layout and construction activities will create variations in vehicle flow within the work zone. To estimate the sample size requirements, it is advisable to use the standard deviation to measure the data dispersion, and a minimum of three initial test runs is required. GPS devices with sufficient accuracy usually require 5 to 10 samples for travel time and delay studies and work zone studies. Stopped delay studies may require a large sample of up to 30 test runs.

scholarly journals Determinants of Target Victim Selection: A Case Study of Criminals from Gambaga Prisons

2020 ◽  
Vol 3 (1) ◽  
pp. 93-112
Author(s):  
Francess Dufie Azumah ◽  
John Onzaberigu Nachinaab ◽  
Samuel Krampah ◽  
Pius Nzeh Ayim
Keyword(s):  
Data Collection ◽  
Sample Size ◽  
Main Tool ◽  
Qualitative Approach ◽  
Sampling Techniques ◽  
Victim Selection ◽  
Interview Guide ◽  
Selection Of

The study was conducted to explore the determinants of target victim selection by criminals, mode of operations adopted by criminals and the factors considered in the selection of victims by criminals. The study adopted a qualitative approach where simple random and purposively sampling techniques were used to select a sample size of 50 inmates in the Gambaga prisons. Interview guide was the main tool used for the data collection for this study. The study found that victimization influences criminal intentions and behaviour of an individual. The inmates noted that they like to attack victims in isolation and in dark places, especially at night. The study further found that criminals operated in gangs, in areas with darkness and in the night using caps and tattered dresses that anonymized them. They also used fear to traumatize their victims, monitor their victims, as well as operated with guns and knives and under special requests of other individuals in society.

scholarly journals Sample Sizes Needed to Describe Length-Frequency of Small-Bodied Fishes: An Example Using Larval Pacific Lamprey

2016 ◽  
Vol 7 (2) ◽  
pp. 315-322 ◽  
Author(s):  
Luke D. Schultz ◽  
Mariah P. Mayfield ◽  
Steven L. Whitlock
Keyword(s):  
Data Collection ◽  
Sample Size ◽  
Length Distribution ◽  
Fish Population ◽  
Anadromous Fish ◽  
Western North America ◽  
Sample Sizes ◽  
Length Frequency ◽  
Pacific Lamprey ◽  
Number Of Individuals

Abstract The ability to describe the length distribution of a fish population requires sampling an adequate number of individuals, but collecting more fish than needed is inefficient. While fisheries managers have assessed sample size requirements for many sport fishes, these requirements are not routinely described for small-bodied fishes (i.e., maximum length ≤200 mm), particularly larval lampreys. To improve the efficiency of data collection for these fishes, we used resampling analyses to asses sample size requirements for accurately describing length distributions of larval (freshwater-dwelling) Pacific lamprey Entosphenus tridentatus, an anadromous fish native to western North America (total length 60–156 mm). We found that the highest increases in accuracy occurred with sample sizes &lt;50, and that we needed sample sizes of 40 to 130 to describe length frequency with 95% confidence, depending on length interval used for performing length-frequency analyses. From these results, we recommend collecting 100 individuals if using 5-mm length intervals to examine length frequency of larval lamprey. These findings can also be used to estimate the relative accuracy of sample sizes in existing assessments and develop and refine monitoring programs for larval lampreys and other small-bodied fishes.

scholarly journals Adaptive sample size determination for the development of clinical prediction models

2021 ◽  
Vol 5 (1) ◽  
Author(s):  
Evangelia Christodoulou ◽  
Maarten van Smeden ◽  
Michael Edlinger ◽  
Dirk Timmerman ◽  
Maria Wanitschek ◽  
...  
Keyword(s):  
Ovarian Cancer ◽  
Sample Size ◽  
Prediction Models ◽  
A Priori ◽  
Sample Size Calculation ◽  
Model Performance ◽  
Sample Size Determination ◽  
Sample Sizes ◽  
Cancer Data ◽  
C Statistic

Abstract Background We suggest an adaptive sample size calculation method for developing clinical prediction models, in which model performance is monitored sequentially as new data comes in. Methods We illustrate the approach using data for the diagnosis of ovarian cancer (n = 5914, 33% event fraction) and obstructive coronary artery disease (CAD; n = 4888, 44% event fraction). We used logistic regression to develop a prediction model consisting only of a priori selected predictors and assumed linear relations for continuous predictors. We mimicked prospective patient recruitment by developing the model on 100 randomly selected patients, and we used bootstrapping to internally validate the model. We sequentially added 50 random new patients until we reached a sample size of 3000 and re-estimated model performance at each step. We examined the required sample size for satisfying the following stopping rule: obtaining a calibration slope ≥ 0.9 and optimism in the c-statistic (or AUC) < = 0.02 at two consecutive sample sizes. This procedure was repeated 500 times. We also investigated the impact of alternative modeling strategies: modeling nonlinear relations for continuous predictors and correcting for bias on the model estimates (Firth’s correction). Results Better discrimination was achieved in the ovarian cancer data (c-statistic 0.9 with 7 predictors) than in the CAD data (c-statistic 0.7 with 11 predictors). Adequate calibration and limited optimism in discrimination was achieved after a median of 450 patients (interquartile range 450–500) for the ovarian cancer data (22 events per parameter (EPP), 20–24) and 850 patients (750–900) for the CAD data (33 EPP, 30–35). A stricter criterion, requiring AUC optimism < = 0.01, was met with a median of 500 (23 EPP) and 1500 (59 EPP) patients, respectively. These sample sizes were much higher than the well-known 10 EPP rule of thumb and slightly higher than a recently published fixed sample size calculation method by Riley et al. Higher sample sizes were required when nonlinear relationships were modeled, and lower sample sizes when Firth’s correction was used. Conclusions Adaptive sample size determination can be a useful supplement to fixed a priori sample size calculations, because it allows to tailor the sample size to the specific prediction modeling context in a dynamic fashion.

scholarly journals Sample Size Determination for the Bayesian t-test and Welch’s test Using the Approximate Adjusted Fractional Bayes Factor

2019 ◽  
Author(s):  
Qianrao Fu ◽  
Herbert Hoijtink ◽  
Mirjam Moerbeek
Keyword(s):  
Sample Size ◽  
Bayes Factor ◽  
Alternative Hypothesis ◽  
Threshold Value ◽  
R Package ◽  
T Test ◽  
Sample Size Determination ◽  
Size Determination ◽  
Fractional Bayes Factor

When two independent means $\mu_1$ and $\mu_2$ are compared, $H_0: \mu_1=\mu_2$, $H_1: \mu_1\ne\mu_2$, and $H_2: \mu_1&gt;\mu_2$ are the hypotheses of interest. This paper introduces the \texttt{R} package \texttt{SSDbain}, which can be used to determine the sample size needed to evaluate these hypotheses using the Approximate Adjusted Fractional Bayes Factor (AAFBF) implemented in the \texttt{R} package \texttt{bain}. Both the Bayesian t-test and the Bayesian Welch's test are available in this \texttt{R} package. The sample size required will be calculated such that the probability that the Bayes factor is larger than a threshold value is at least $\eta$ if either the null or alternative hypothesis is true. Using the \texttt{R} package \texttt{SSDbain} and/or the tables provided in this paper, psychological researchers can easily determine the required sample size for their experiments.

scholarly journals When “Normal” Becomes Normative: A Case Study of Researchers’ Quotation Errors When Referring to a Focus Group Sample Size Study

2019 ◽  
Vol 18 ◽  
pp. 160940691984125 ◽  
Author(s):  
Claire Glenton ◽  
Benedicte Carlsen
Keyword(s):  
Focus Group ◽  
Sample Size ◽  
Google Scholar ◽  
Sample Sizes ◽  
Typical Sample ◽  
Descriptive Information ◽  
Very High

In 2011, we published a review exploring how researchers report and justify their focus group sample sizes. We concluded that sample sizes vary widely and that most researchers give no explanation for their sample size. The aim of our 2011 study was to describe practice rather than develop guidance. However, after our study was published, we noticed that new researchers were using our information about typical sample sizes as justification for their own sample size. In other words, practice that we had presented as typical or “normal” but generally lacking in justification was being used as normative. The current study aims to explore the misrepresentation of descriptive information as normative. Specifically, we map this type of quotation error in references to our 2011 study. Using Google Scholar, we identified all articles referencing our study. We then extracted quotations where the researchers had referred to our study and categorized these as follows: (a) quotations where the researchers had used the descriptive information from our study to justify their sample size and (b) quotations where the researchers had referred to our study for other purposes or where the purpose was unclear. We assessed 205 articles that had referred to our 2011 study. We identified the type of quotation error we were interested in, namely the misrepresentation of descriptive information as normative, in 50.7% of the included articles. Our study shows very high rates of one type of quotation error: the misrepresentation of descriptive information about focus group sample size as normative. Researchers referring to other researchers’ work carry most of the responsibility for ensuring that they do this appropriately. However, the authors of the research being referred to also need to consider how they can make their results clearer. We offer suggestions as to how this might be achieved.

Sign in / Sign up

Export Citation Format

Share Document