元分析中的方法问题及常用英文表达

目录

1.      元分析/文献综述

1.1       文献搜集

1.2       初步编码/二次编码

1.3       纳入/排除标准

1.4       数据提取

1.5       研究质量评估Study quality assessment

1.6       如何处理重复发表问题How dealt with replicated publications

1.7       编码一致性Coding agreement

1.7.1       Cohen’s Kappa

1.8       调节变量Moderations

2.      参考文献

1.               元分析/文献综述

1.1             文献搜集

An extensive search was conducted to identify asmany published and unpublished studies as possible. The process involved asearch of computerized databases from 1990 to April of 2010. Databases utilizedin the search included the following: ….Web of Science etc.

The search included the terms xxxxx,We also conducted a manual search ofmajor journals (e.g., JAP AMJ, JOB…) as well as the reference lists ofpertinent articles on work engagement. Finally, we collected unpublisheddissertations, conference presentations, and emailed authors of publishedresearch on X to obtain any unpublished work. This process resulted in over Npublished and over N2 unpublished articles (Christian,Garza, & Slaughter, 2011).

1.2             FirstCoding/second coding

We documented the following information foridentification and quality assessment of the study, and the statisticalanalyses afterwards: bibliographic information, sector of the sample, samplesize, effect sizes between relevant indicators, study design, number of andinterval between measurements, psychometric quality of the measures, method ofanalysis, and response rate and nonresponse analysis (Lesener,Gusy, & Wolter, 2018).

1.3             Inclusion/exclusioncriteria

To be included, studies had to meet the followingfour criteria: The study (1) investigated employees based on the m model; (2)used a longitudinal design with two or more waves; (3) measured at least onejob characteristic (X or Y), and (4) a dimension of well-being (e.g., workengagement or burnout) (Leseneret al., 2018).

1.4             Data extraction

1)Characteristics of the studies 

2).Demographic information 

3)Study characteristicsincluded the author details and document date, type of document found (published article, thesis, or grey literature), country of interventionlocation, the core components of theintervention (e.g. workshops, coaching, assignments), intervention type (asdefined in the introduction) and intervention style (whether or not theintervention was group or individually orientated, a combined group andindividual intervention, or conducted purely online). 

1.5             Study quality assessment

For quality assessment, we used a scheme from DeLange et al. (2003), which was developed to evaluate the quality of studies inoccupational health psychology. 

It consists of five criteria: 

(1) study design,

(2) number and interval of measurements, 

(3) psychometric quality of themeasures, 

(4) statistical analyses, and 

(5) nonresponse analysis. Studies couldobtain between one (insufficient) and four (very good) stars for each criterion.

Table 1 summarises the criteria and ratings.

Figures (Lesener, Gusy, & Wolter, 2018)

1.6             How  dealt with replicated publications

Study effects developed from the same data set andduplicated in more than one publication likely will bias the aggregated effectsin a meta-analysis. The following heuristic, as outlined in Figure 1, developsa system for the detection of potential duplicate studies (Wood,2008).

 Wood, J. (2008). 

1.7             Coding agreement

1.7.1           Cohen’sKappa

reviewers sometime may ask you to provide Kappa “Report Cohen’s kappa instead of percentagreement as it takes into consideration the possibility for agreement to occurby chance. Also, what was inter-rater reliability for the sample size,correlation, and alphas?”

 I find awebsite you can just input the data and get your results!

https://www.graphpad.com/quickcalcs/kappa1/

1.8             Moderations

To test for homogeneity in the population effect size estimate, we usedSchmidt and Hunter’s (2015) 75% rule. This test is a rule of thumb fordetermining if the heterogeneity is high enough to be meaningful. The ratio of the estimate of error variance in the population effect size to variance in theobserved effect sizes was compared with 0.75. If this value exceeds 0.75, the remaining unexplained variance is likely due to uncorrected artifacts in thestudies and can be ignored (Kimet al., 2018).

2.               References

Christian, M. S., Garza, A. S., & Slaughter, J. E. (2011). WORKENGAGEMENT: A QUANTITATIVE REVIEW AND TEST OF ITS RELATIONS WITH TASK ANDCONTEXTUAL PERFORMANCE. PERSONNEL PSYCHOLOGY (Vol. 64). Retrievedfrom https://onlinelibrary.wiley.com/doi/pdf/10.1111/j.1744-6570.2010.01203.x

Judge, T. A., &Piccolo, R. F. (2004). Transformational and Transactional Leadership: AMeta-Analytic Test of Their Relative Validity. Journal of Applied Psychology,89(5), 755–768. https://doi.org/10.1037/0021-9010.89.5.755

Kim, M., Beehr, T.A., & Prewett, M. S. (2018). Employee Responses to Empowering Leadership: AMeta-Analysis. Journal of Leadership & Organizational Studies,154805181775053. https://doi.org/10.1177/1548051817750538

Knight, C.,Patterson, M., & Dawson, J. (2017). Building work engagement: A systematicreview and meta-analysis investigating the effectiveness of work engagementinterventions. Journal of Organizational Behavior, 38(6), 792–812.https://doi.org/10.1002/job.2167

Lesener, T., Gusy,B., & Wolter, C. (2018). The job demands-resources model: A meta-analyticreview of longitudinal studies. Work & Stress.https://doi.org/10.1080/02678373.2018.1529065

Rabl, T., Jayasinghe,M., Gerhart, B., & Kühlmann, T. M. (2014). A meta-analysis of countrydifferences in the high-performance work system-business performancerelationship: The roles of national culture and managerial discretion. Journalof Applied Psychology, 99(6), 1011–1041. https://doi.org/10.1037/a0037712

Wood, J. (2008).Methodology for dealing with duplicate study effects in a meta-analysis. OrganizationalResearch Methods, 11(1), 79–95.https://doi.org/10.1177/1094428106296638

完整笔记pdf版本可以通过点击read more下载 （7天内有效）