運用職業重建資訊系統資料進行身心障礙服務個案就業之空間分析
No Thumbnail Available
Date
2021-11-??
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
國立台灣師範大學特殊教育學系
National Taiwan Normal University Department of Special Education
National Taiwan Normal University Department of Special Education
Abstract
為促進身心障礙者就業，許多研究已致力於就業成效的評估與各項影響因素的探討，但是對於身心障礙者就業之空間位置與地理特性分析，迄今仍甚少涉及。近年來，空間分析方法迅速進步，有助了解地物分布所呈現的空間型態，並可結合空間迴歸模型，更確切掌握區域特性與各項因素間的關係。因此本研究旨在運用民國105 年「全國身心障礙者職業重建個案服務資訊管理系統」內成功就業個案的資料，結合空間分析方法，探討臺灣接受職業重建服務之身心障礙個案的就業分布狀況，共計使用4,592 筆身心障礙個案的就業數據。分析過程依序包括空間自相關分析、相關分析與迴歸分析，以ArcGIS 10.2 軟體，進行空間資料分析，並利用SPSS Statistics 17.0，進行統計數據分析。結果發現，身心障礙個案就業分布經全域型空間自相關Moran’s Index 檢定顯示呈現空間群聚現象；區域型空間自相關Local Indicator of Spatial Association (LISA) 分析得知，身心障礙個案就業群聚的現象以高屬性群聚（High-High）為主，未有顯著的高屬性離群（High-Low）、低屬性群聚（Low-High）與低屬性離群（Low-Low）區域發現。迴歸分析顯示，從業人口密度與服務業佔比是影響身心障礙個案就業之顯著因子，具有顯著正相關，且這兩個變數在不同地區產生不同的影響性；地理加權迴歸模型解釋力達86%。依據研究結果，本研究認為空間分析在職業重建領域的應用需要持續被重視，使用地理加權迴歸較傳統線性迴歸更能解釋臺灣接受職業重建服務個案的就業分布狀況，區域空間分析結果亦可提供更細緻的實務應用建議。
Purpose: To facilitate the employment of people with disabilities, many studies have evaluated employment outcomes and investigated the factors relevant to these outcomes. However, studies investigating the geospatial distribution of disability employmentremain scarce. Rapid development in spatial analysis has helped researchers understand spatial distribution patterns. Spatial regression models can be used to investigate the relationships between various factors potentially associated with spatial characteristics. Therefore, this study applied spatial analysis to explore the spatial distribution patterns of the employment of clients with disabilities who were successfully employed after they received vocational rehabilitation services in Taiwan. Methods: Employment data were collected from the National Disability Vocational Rehabilitation Case Service Database. There were 4,592 clients who were engaged in paid employment after receiving the services in 2016. Data analysis included spatial autocorrelation analysis, correlation analysis, and regression analysis. In the spatial autocorrelation analysis, two indicators were used: Moran's Index measured global spatial autocorrelation based on both feature locations and feature values simultaneously to explore an overall spatial distribution pattern, while local indicators of spatial association (LISA) assessed the possibility of recognition of spatial clusters in each local data sets and the spatial patterns of the indictors were categorized into four zones (high-high, low-low, high-low and lowhigh). In the regression analysis, the traditional ordinary least-squares regression was applied first. It was then followed by the geographically weighted regression due to the identification of spatial autocorrelation in residuals. The ArcGIS 10.2 and SPSS Statistics 17.0 software packages were used to conduct the spatial and statistical analyses, respectively. Results/Findings: Global spatial autocorrelation analysis indicated spatial clusters in the employment of clients with disabilities. A significant high–high pattern was identified through local spatial autocorrelation analysis using local indicators of spatial association, but high–low, low–high, and low–low patterns were not identified. Moreover, the regression analysis indicated that employment density and service industry percentage were predictors of the geospatial distribution of the employment of clients with disabilities. These two variables were positively correlated and exhibited varied effects in different townships in Taiwan. The geographically weighted regression model accounted for 86% of the variance in the geospatial distribution of disability employment. Conclusions/Implications: The results give evidence of the importance of using spatial analysis in the vocational rehabilitation field. More endeavors are needed to increase the knowledge. The geographically weighted regression has the potential to provide a more accurate result than the traditional ordinary least-squares regression in determining the spatial distribution of employment of clients with disabilities who received vocational rehabilitation services in Taiwan. Further implications for practice based on the local spatial distribution patterns identified can also be provided herein.
Purpose: To facilitate the employment of people with disabilities, many studies have evaluated employment outcomes and investigated the factors relevant to these outcomes. However, studies investigating the geospatial distribution of disability employmentremain scarce. Rapid development in spatial analysis has helped researchers understand spatial distribution patterns. Spatial regression models can be used to investigate the relationships between various factors potentially associated with spatial characteristics. Therefore, this study applied spatial analysis to explore the spatial distribution patterns of the employment of clients with disabilities who were successfully employed after they received vocational rehabilitation services in Taiwan. Methods: Employment data were collected from the National Disability Vocational Rehabilitation Case Service Database. There were 4,592 clients who were engaged in paid employment after receiving the services in 2016. Data analysis included spatial autocorrelation analysis, correlation analysis, and regression analysis. In the spatial autocorrelation analysis, two indicators were used: Moran's Index measured global spatial autocorrelation based on both feature locations and feature values simultaneously to explore an overall spatial distribution pattern, while local indicators of spatial association (LISA) assessed the possibility of recognition of spatial clusters in each local data sets and the spatial patterns of the indictors were categorized into four zones (high-high, low-low, high-low and lowhigh). In the regression analysis, the traditional ordinary least-squares regression was applied first. It was then followed by the geographically weighted regression due to the identification of spatial autocorrelation in residuals. The ArcGIS 10.2 and SPSS Statistics 17.0 software packages were used to conduct the spatial and statistical analyses, respectively. Results/Findings: Global spatial autocorrelation analysis indicated spatial clusters in the employment of clients with disabilities. A significant high–high pattern was identified through local spatial autocorrelation analysis using local indicators of spatial association, but high–low, low–high, and low–low patterns were not identified. Moreover, the regression analysis indicated that employment density and service industry percentage were predictors of the geospatial distribution of the employment of clients with disabilities. These two variables were positively correlated and exhibited varied effects in different townships in Taiwan. The geographically weighted regression model accounted for 86% of the variance in the geospatial distribution of disability employment. Conclusions/Implications: The results give evidence of the importance of using spatial analysis in the vocational rehabilitation field. More endeavors are needed to increase the knowledge. The geographically weighted regression has the potential to provide a more accurate result than the traditional ordinary least-squares regression in determining the spatial distribution of employment of clients with disabilities who received vocational rehabilitation services in Taiwan. Further implications for practice based on the local spatial distribution patterns identified can also be provided herein.