Estimating the Stability of Psychological Dimensions via Bootstrap Exploratory Graph Analysis: A Monte Carlo Simulation and Tutorial

Exploratory Graph Analysis (EGA) has emerged as a popular approach for estimating the dimensionality of multivariate data using psychometric networks. Sampling variability, however, has made reproducibility and generalizability a key issue in network psychometrics. To address this issue, we have developed a novel bootstrap approach called Bootstrap Exploratory Graph Analysis (bootEGA). bootEGA generates a sampling distribution of EGA results where several statistics can be computed. Descriptive statistics (median, standard error, and dimension frequency) provide researchers with a general sense of the stability of their empirical EGA dimensions. Structural consistency estimates how often dimensions are replicated exactly across the bootstrap replicates. Item stability statistics provide information about whether dimensions are unstable due to misallocation (e.g., item placed in the wrong dimension), multidimensionality (e.g., item belonging to more than one dimension), and item redundancy (e.g., similar semantic content). Using a Monte Carlo simulation, we determine guidelines for acceptable item stability. After, we provide an empirical example that demonstrates how bootEGA can be used to identify structural consistency issues (including a fully reproducible R tutorial). In sum, we demonstrate that bootEGA is a robust approach for identifying the stability and robustness of dimensionality in multivariate data.

Grain Yield Performance and Parametric Stability Statistics of Tef {Eragrostis tef (Zucc) Trotter} Genotypes in Tigray, Ethiopia

The most constraints of tef productions are lodging, drought, low yield cultivars; insect and disease affected the growth of tef. These, factors causes inconsistence performance yield due to GEI. The objective was to evaluate tef genotypes on their yield performance, stability and parametric stability to select most independent and informative statistics method. The experiment was conducted at four locations for two seasons; with design of RCBD three replications, two standard checks and 19 tef genotypes. Data was collected on grain yield and analyzed by R software and STABILITYSOFT. The analysis of variance for the combined mean of grain yield showed that there was significance difference (P<0.001) between genotypes, environments and GEI. Yield performance was influenced by Environments and GEI. The mean grain yield of genotypes over GEI varies from 820.94kg/ha to 2438.90kg/ha, while the genotype grain yield was ranged from 1382 to 1989kg/ha. G19, G17 and G6 were identified the higher grain yield performance over seven environments. Whereas, G8 and G11 were the lowest yielding tef genotypes. Nine parametric methods and GGE biplot were used to evaluate the stability of the genotypes. G19 was the most stable following G17 and would be grown for unfavorable growing environments. However, G6 was stable for favorable environmental condition. G19 and G17 had static stability and fitting for area faced with erratic rain fall. Even though, parametric stability did not show a positive and statistically significant correlation with mean yield the Mean variance component (θi) is selected with GGE biplot for evaluation of tef genotypes in the development of cultivar. Effective selection of variety would be best if mega-environment, representative and discriminating testing areas are identified.

Correlations Between Stability Statistics of Forage Production in Elephant Grass

Elephant grass (Pennisetum purpureum Schum.) is an important forage plant in the tropics and the potential of genotypes depends on the genotype &times; environment interaction effects. The objective of this study was to evaluate and compare different stability methods of forage production of 53 elephant grass genotypes, in Campos dos Goytacazes, Rio de Janeiro State, Brazil. The experiment lasted two years, a total of ten cuts with randomized block experimental design with two replications. The analysis of variance was applied to data from dry matter production (DMP), subjected to stability analysis using the following methods: Yates and Cochran, Plaisted and Peterson, ecovalence Wrickie, Kang and Phan, Lin and Bins, and Annicchiarico. The Yates and Cochran method showed more stable genotypes but being less productive. Plaisted and Peterson and ecovalence Wrickie methods presented a Spearman correlation equal to 1, so it is not recommended to implement them concurrently. Lin and Bins showed a strong negative correlation with the average being a method that indicates the genotype also very stable and productive. This method correlates with Annicchiarico, which also indicates productive genotypes by the confidence index. The genotypes most stable among the methods were: Pusa Napier 2, Taiwan A-143 and Merckeron Comum.

Comparison of Non-parametric Stability Statistics for Selecting Stable and Adapted Soybean Genotypes under Different Environments

The objectives of this study were to investigate the comparison among non-parametric stability statistics and to evaluate seed yield stability of the sixteen soybean genotypes across four locations during the 2016, 2017 and 2018 growing seasons in Egypt. All trials were laid down in a randomized complete block design (RCBD) with three replications. The AMMI analysis showed ahighly significant effect of genotype (G), environment (E) and G x E interaction (GEI). The major contributions to treatment sum of squares were GEI, followed by G and E. The AMMI analysis also partitioned the total GEI component into eleven PCAs and Residual. The first eight PCAs were highly significant and accounted for about 99.56% of the total GEI. Based on the static and dynamic concepts, the results of spearman’s rank correlation and PCA showed that stability measures could be classified into three groups. The non-parametric stability statistics i.e., YSi, KR, TOP, RSM and δgy related to the dynamic concept and strongly correlated with mean seed soybean yield of stability. While, the other non-parametric stability statistics (Si(1) ,Si(2) ,Si(3) and Si(6),NPi(1) ,NPi(2) ,NPi(3) and NPi(4) , δr, MID, LOW) represented the concept of static stability, which were influenced simultaneously by both yield and stability. The non-parametric stability statistics in each the groups I, II, and III were positively and significantly correlated with each other, thus; any of these parameters could be considered as appropriate alternatives for each other. According to cluster analysis, soybean genotypes G6, G4, G8, G11, G9, G1, G7 and G2 were more stable varieties on the basis of mean seed yield and non-parametric stability statistics. In conclusion, both yield and stability should be considered simultaneously to exploit the useful effect of GEI and to make the selection of genotypes more precise and refined. Thus, the YSi, KR, TOP, RSM and δgy were more useful statistics in soybean breeding programmes and could be useful alternatives to parametric stability statistics. According to most non-parametric stability statistics, the genotypes G6 and G11 were more stable coupled with high seed yield; therefore, these genotypes might be used for genetic improvement of soybean and they must be released in studied regions and other regions in Egypt.

Estimating the stability of the number of factors via Bootstrap Exploratory Graph Analysis: A tutorial

Exploratory graph analysis (EGA) is a recent addition to the growing field of network psychometrics. EGA has emerged as a popular approach for estimating the dimensonality of data in networks. The appeal of EGA is the visualization of the relations between variables and the deterministic allocation of variables into dimensions. Notably, networks tend to be sample-specific, making reproducibility and generalizability a key issue in network psychometrics. To resolve this issue, we’ve developed a novel bootstrap approach called, Bootstrap Exploratory Graph Analysis (bootEGA). bootEGA provides researchers with dimension and item stability statistics as well as item analyses that are akin to exploratory factor analysis loadings. We provide descriptions of bootEGA’s functions accompanied by a step-by-step R tutorial for how to apply and interpret bootEGA’s results. This tutorial is applied to real-world data to demonstrate its effectiveness at identifying problematic dimensions and items. In short, our results show that bootEGA is a robust approach for identifying the stability and robustness of dimensionality in data.