Imputation of missing links and attributes in longitudinal social surveys

Ouzienko, Vladimir; Obradović, Zoran

doi:10.1007/s10994-013-5420-1

Cited by 12 publications

(4 citation statements)

References 47 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Third, strategic transactions are a meaningful measure of firms structural embeddedness as confirmed by the literature review in the theoretical framework section. Specifically, strategic alliances which make up 74.5 percent of all strategic transactions in our data have long been considered an optimal source for centrality measures' analysis [73]. Fourth, there is a lack of studies on dynamic embeddedness applied to the global pharmaceutical industry.…”

Section: Empirical Analysis 31 Research Contextmentioning

confidence: 99%

Organizational dynamic embeddedness and external shocks: The impact of financial and recession crises in strategic networks of the global pharmaceutical industry

2016

View full text Add to dashboard Cite

We analyze the global pharmaceutical industry network using a unique database that covers strategic transactions (i.e., alliance, financing and acquisition collaborations) for the top 90 global pharmaceutical firms and their ego-network partnerships totaling 4735 members during 1991À2012. The article explores insights on dynamic embeddedness analysis under network perturbations by exploring core and full networks' behavior during the global financial crisis of 2007À2008 and the subsequent global and Eurozone recessions of 2009À2012. We introduce and test literature grounded hypotheses as well as report network visualizations and nonparametric tests that reveal important discrepancies in both network types before and after the financial crisis offset. We observe that firms in core and full networks behave differently, with smaller top pharmaceutical firms of core networks particularly being affected by the crises, potentially due to a collaboration reduction with bigger top pharmaceuticals. On the other hand, big pharmaceuticals in full networks maintain their centrality position as a possible consequence of their strategic collaborations not only with other similarly sized firms but also due to their connections with subsidiaries and other private entities present in the total sample. Our results confirm the significant dynamicity reduction during financial crisis and recession periods for core and full networks, and highlight the importance that exogenous factors as well as network types play in centrality-based dynamic longitudinal network analysis.

show abstract

Section: Empirical Analysis 31 Research Contextmentioning

confidence: 99%

Organizational dynamic embeddedness and external shocks: The impact of financial and recession crises in strategic networks of the global pharmaceutical industry

2016

View full text Add to dashboard Cite

show abstract

“…Constrained Random Dot Product Graph (DP) models actors in an s-dimensional latent space, and the missing edges are imputed by the dot product of each pair of actors' latent position vectors (Marchette and Priebe, 2008;Ouzienko and Obradovic, 2014). The imputation process can be described as the following.…”

Section: Constrained Random Dot Product Graphmentioning

confidence: 99%

Comparison of Methods for Imputing Social Network Data

Xu¹,

Hai²,

Yang³

et al. 2022

Journal of Data Science

View full text Add to dashboard Cite

Social network data often contain missing values because of the sensitive nature of the information collected and the dependency among the network actors. As a response, network imputation methods including simple ones constructed from network structural characteristics and more complicated model-based ones have been developed. Although past studies have explored the influence of missing data on social networks and the effectiveness of imputation procedures in many missing data conditions, the current study aims to evaluate a more extensive set of eight network imputation techniques (i.e., null-tie, Reconstruction, Preferential Attachment, Constrained Random Dot Product Graph, Multiple Imputation by Bayesian Exponential Random Graph Models or BERGMs, k-Nearest Neighbors, Random Forest, and Multiple Imputation by Chained Equations) under more practical conditions through comprehensive simulation. A factorial design for missing data conditions is adopted with factors including missing data types, missing data mechanisms, and missing data proportions, which are applied to generated social networks with varying numbers of actors based on 4 different sets of coefficients in ERGMs. Results show that the effectiveness of imputation methods differs by missing data types, missing data mechanisms, the evaluation criteria used, and the complexity of the social networks. More complex methods such as the BERGMs have consistently good performances in recovering missing edges that should have been present. While simpler methods like Reconstruction work better in recovering network statistics when the missing proportion of present edges is low, the BERGMs work better when more present edges are missing. The BERGMs also work well in recovering ERGM coefficients when the networks are complex and the missing data type is actor non-response. In conclusion, researchers analyzing social networks with incomplete data should identify the network structures of interest and the potential missing data types before selecting appropriate imputation methods.

show abstract

“…In these models, expected values are imputed for missing data during the estimation of the ERGM parameters; because ERGMs are estimated via simulation (usually employing standard Markov chain Monte Carlo methods), the missing data are imputed many times. While still quite new and rare, these ERGM-based procedures have been successfully implemented on real-world network data suffering from quite complicated patterns of missingness (Desmarais and Cranmer 2012;,Wang et al 2016), and have been mathematically extended to temporal ERGMs (Leifeld et al 2017;Ouzienko and Obradovic 2013). We need more validation to understand fully the properties and reliability of these model-based methods, but they possess considerable promise for scholars analysing network data containing missing information.…”

Section: Addressing Data Completenessmentioning

confidence: 99%