Assessing the Potential of Multiple Imputation in Sequence Analysis

Sequence analysis and optimal matching are useful heuristic tools for the descriptive analysis of heterogeneous individual pathways such as educational careers, job sequences or patterns of family formation. However, to date it remains unclear how to handle the inevitable problems caused by missing values with regard to such analysis. Multiple Imputation (MI) offers a possible solution for this problem but it has not been tested in the context of sequence analysis. Against this background, we contribute to the literature by assessing the potential of MI in the context of sequence analyses using an empirical example. Methodologically, we draw upon the work of Brendan Halpin and extend it to additional types of missing value patterns. Our empirical case is a sequence analysis of panel data with substantial attrition that examines the typical patterns and the persistence of sex segregation in school-to-work transitions in Switzerland. The preliminary results indicate that MI is a valuable methodology for handling missing values due to panel mortality in the context of sequence analysis. MI is especially useful in facilitating a sound interpretation of the resulting sequence types.

A 49 year hindcast of surface winds over the Iberian Peninsula

The use of hindcast climatic data is quite extended for multiple applications. However, this approach needs the support of a validation process to allow its drawbacks and, therefore, confidence levels to be assessed. In this work, the strategy relies on an hourly wind database resulting from a dynamical downscaling experiment, with a spatial resolution of 10 km, covering the Iberian Peninsula (IP), driven by the ERA40 reanalysis (1959–2001) extended by European Centre for Medium-Range Weather Forecast (ECMWF) analysis (2002–2007) and comprising two main steps. Initially, the skill of the simulation is evaluated comparing the quality-tested observational database (Lorente-Plazas et al., 2014) at local and regional scales. The results show that the model is able to portray the main features of the wind over the IP: annual cycles, wind roses, spatial and temporal variability, as well as the response to different circulation types. In addition, there is a significant added value of the simulation with respect to driving conditions, especially in regions with a complex orography. However, some problems are evident, the major drawback being the systematic overestimation of the wind speed, which is mainly attributed to a missrepresentation of frictional forces. The model skill is also lower along the Mediterranean coast and for the Pyrenees. In a second phase, the high spatio-temporal resolution of the pseudo-real wind database is used to explore the limitations of the observational database. It is shown that missing values do not affect the characterisation of the wind climate over the IP, while the length of the observational period (6 years) is sufficient for most regions, with only a few exceptions. The spatial distribution of the observational sampling schemes should be enhanced to improve the correct assessment of all IP wind regimes, particularly in some mountainous areas.

Challenges in Collating Spirometry Reference Data for South-Asian Children: An Observational Study.

METHODS Spirometry datasets from South-Asian children were collated from four centres in India and five within the UK. Records with transcription errors, missing values for height or spirometry, and implausible values were excluded(n = 110). RESULTS Following exclusions, cross-sectional data were available from 8,124 children (56.3% male; 5-17 years). When compared with GLI-predicted values from White Europeans, forced expired volume in 1s (FEV1) and forced vital capacity (FVC) in South-Asian children were on average 15% lower, ranging from 4-19% between centres. By contrast, proportional reductions in FEV1 and FVC within all but two datasets meant that the FEV1/FVC ratio remained independent of ethnicity. The 'GLI-Other' equation fitted data from North India reasonably well while 'GLI-Black' equations provided a better approximation for South-Asian data than the 'GLI-White' equation. However, marked discrepancies in the mean lung function z-scores between centres especially when examined according to socio-economic conditions precluded derivation of a single South-Asian GLI-adjustment. CONCLUSION Until improved and more robust prediction equations can be derived, we recommend the use of 'GLI-Black' equations for interpreting most South-Asian data, although 'GLI-Other' may be more appropriate for North Indian data. Prospective data collection using standardised protocols to explore potential sources of variation due to socio-economic circumstances, secular changes in growth/predictors of lung function and ethnicities within the South-Asian classification are urgently required.

Logistic regression with incompletely observed binary

Logistic regression is one of the most important tools in the analysis of epidemiological and clinical data. Such data often contain missing values for one or more variables. Common practice is to eliminate all individuals for whom any information is missing. This deletion approach does not make efficient use of available information and often introduces bias.^ Two methods were developed to estimate logistic regression coefficients for mixed dichotomous and continuous covariates including partially observed binary covariates. The data were assumed missing at random (MAR). One method (PD) used predictive distribution as weight to calculate the average of the logistic regressions performing on all possible values of missing observations, and the second method (RS) used a variant of resampling technique. Additional seven methods were compared with these two approaches in a simulation study. They are: (1) Analysis based on only the complete cases, (2) Substituting the mean of the observed values for the missing value, (3) An imputation technique based on the proportions of observed data, (4) Regressing the partially observed covariates on the remaining continuous covariates, (5) Regressing the partially observed covariates on the remaining continuous covariates conditional on response variable, (6) Regressing the partially observed covariates on the remaining continuous covariates and response variable, and (7) EM algorithm. Both proposed methods showed smaller standard errors (s.e.) for the coefficient involving the partially observed covariate and for the other coefficients as well. However, both methods, especially PD, are computationally demanding; thus for analysis of large data sets with partially observed covariates, further refinement of these approaches is needed. ^

Geochemistry of ODP Site 125-786 volcanic rocks

Eocene-Oligocene volcanic rocks drilled at Site 786 in the Izu-Bonin forearc cover a wide range of compositions from primitive boninites to highly evolved rhyolites. K-Ar dating reveals at least two distinct episodes of magmatism; one at 41 Ma and a later one at 35 Ma. The early episode produced low-Ca boninites and bronzite andesites that form an oceanic basement of pillow lavas and composite intrusive sheets, overlain by flows and intrusive sheets of intermediate-Ca boninites and bronzite-andesites and a fractionated series of andesites, dacites, and rhyolites. The later episode produced high-Ca boninites and intermediate-Ca boninites, exclusively as intrusive sheets.

Acoustic multi-frequency indicator of four major groups on a spatial grid in the Bay of Biscay

The first data set contains the mean and cofficient of variation (standard deviation divided by mean) of a multi-frequency indicator I derived from ER60 acoustic information collected at five frequencies (18, 38, 70, 120, and 200 kHz) in the Bay of Biscay in May of the years 2006, 2008, 2009 and 2010 (Pelgas surveys). The multi-frequency indicator was first calculated per voxel (20 m long × 5 m deep sampling unit) and then averaged on a spatial grid (approx. 20 nm × 20 nm) for five 5-m depth layers in the surface waters (10-15m, 15-20m, 20-25m, 25-30m below sea surface); there are missing values in particular in the shallowest layer. The second data set provides for each grid cell and depth layer the proportion of voxels for which the multi-frequency indicator I was indicative of a certain group of organisms. For this the following interpretation was used: I < 0.39 swim bladder fish or large gas bubbles, I = 0.39-0.58 small resonant bubbles present in gas bearing organisms such as larval fish and phytoplankton, I = 0.7-0.8 fluidlike zooplankton such as copepods and euphausiids, and I > 0.8 mackerel. These proportions can be interpreted as a relative abundance index for each of the four organism groups.

Snow height on sea ice and sea ice drift from autonomous measurements from buoy 2015S22, deployed during the Norwegian Young sea ICE cruise (N-ICE 2015) project

Snow height was measured by the Snow Depth Buoy 2015S22, an autonomous platform, drifting on Arctic sea ice, deployed during the Norwegian Young sea ICE cruise (N-ICE 2015) project. The resulting time series describes the evolution of snow depth as a function of place and time between 2015-03-01 and 2015-05-06 in sample intervals of 1 hour. The Snow Depth Buoy consists of four independent sonar measurements representing the area (approx. 10 m**2) around the buoy. The buoy was installed on first year ice. In addition to snow depth, geographic position (GPS), barometric pressure, air temperature, and ice surface temperature were measured. Negative values of snow depth occur if surface ablation continues into the sea ice. Thus, these measurements describe the position of the sea ice surface relative to the original snow-ice interface. Differences between single sensors indicate small-scale variability of the snow pack around the buoy. The data set has been processed, including the removal of obvious inconsistencies (missing values). Records without any snow depth may still be used for sea ice drift analyses.

Snow height on sea ice and sea ice drift from autonomous measurements from buoy 2015S26, deployed during the Norwegian Young sea ICE cruise (N-ICE 2015) project

Snow height was measured by the Snow Depth Buoy 2015S26, an autonomous platform, drifting on Arctic sea ice, deployed during the Norwegian Young sea ICE cruise (N-ICE 2015) project. The resulting time series describes the evolution of snow depth as a function of place and time between 2015-01-24 and 2015-02-21 in sample intervals of 1 hour. The Snow Depth Buoy consists of four independent sonar measurements representing the area (approx. 10 m**2) around the buoy. The buoy was installed on first year ice. In addition to snow depth, geographic position (GPS), barometric pressure, air temperature, and ice surface temperature were measured. Negative values of snow depth occur if surface ablation continues into the sea ice. Thus, these measurements describe the position of the sea ice surface relative to the original snow-ice interface. Differences between single sensors indicate small-scale variability of the snow pack around the buoy. The data set has been processed, including the removal of obvious inconsistencies (missing values). Records without any snow depth may still be used for sea ice drift analyses.

Snow height on sea ice and sea ice drift from autonomous measurements from buoy 2015S18, deployed during POLARSTERN cruise ANT-XXX/2 (PS89)

Snow height was measured by the Snow Depth Buoy 2015S18, an autonomous platform, drifting on Antarctic sea ice, deployed during POLARSTERN cruise ANT-XXX/2 (PS89). The resulting time series describes the evolution of snow depth as a function of place and time between 2015-01-03 and 2015-01-18 in sample intervals of 1 hour. The Snow Depth Buoy consists of four independent sonar measurements representing the area (approx. 10 m**2) around the buoy. The buoy was installed on first year ice. In addition to snow depth, geographic position (GPS), barometric pressure, air temperature, and ice surface temperature were measured. Negative values of snow depth occur if surface ablation continues into the sea ice. Thus, these measurements describe the position of the sea ice surface relative to the original snow-ice interface. Differences between single sensors indicate small-scale variability of the snow pack around the buoy. The data set has been processed, including the removal of obvious inconsistencies (missing values). Records without any snow depth may still be used for sea ice drift analyses.

Snow height on sea ice and sea ice drift from autonomous measurements from buoy 2014S13, deployed during the CryoVEx2014 field campaign

Snow height was measured by the Snow Depth Buoy 2014S13, an autonomous platform, drifting on Arctic sea ice, deployed during the CryoVEx2014 field campaign. The resulting time series describes the evolution of snow height as a function of place and time between 2014-03-30 and 2014-07-20 in sample intervals of 1 hour. The Snow Depth Buoy consists of four independent sonar measurements representing the area (approx. 10 m**2) around the buoy. The buoy was installed on multi year ice. In addition to snow height, geographic position (GPS), barometric pressure, air temperature, and ice surface temperature were measured. Negative values of snow height occur if surface ablation continues into the sea ice. Thus, these measurements describe the position of the sea ice surface relative to the original snow-ice interface. Differences between single sensors indicate small-scale variability of the snow pack around the buoy. The data set has been processed, including the removal of obvious inconsistencies (missing values). Records without any snow height may still be used for sea ice drift analyses. Note: This data set contains only relative changes in snow height, because no initial readings of absolute snow height are available.

Snow height on sea ice and sea ice drift from autonomous measurements from buoy 2013S1, deployed during Antarctic Fast Ice Network 2013 (AFIN 2013)

Snow height was measured by the Snow Depth Buoy 2013S1, an autonomous platform, installed close to Neumayer III Base, Antarctic during Antarctic Fast Ice Network 2013 (AFIN 2013). The resulting time series describes the evolution of snow height as a function of place and time between 2013-02-11 and 2013-04-29 in sample intervals of 1 hour. The Snow Depth Buoy consists of four independent sonar measurements representing the area (approx. 10 m**2) around the buoy. The buoy was installed on the ice shelf. In addition to snow height, geographic position (GPS), barometric pressure, air temperature, and ice surface temperature were measured. Negative values of snow height occur if surface ablation continues into the sea ice. Thus, these measurements describe the position of the sea ice surface relative to the original snow-ice interface. Differences between single sensors indicate small-scale variability of the snow pack around the buoy. The data set has been processed, including the removal of obvious inconsistencies (missing values). Records without any snow height may still be used for sea ice drift analyses. Note: This data set contains only relative changes in snow height, because no initial readings of absolute snow height are available.

Snow height on sea ice and sea ice drift from autonomous measurements from buoy 2013S3, deployed at the Barneo ice camp 2013

Snow height was measured by the Snow Depth Buoy 2013S3, an autonomous platform, drifting on Arctic sea ice. This buoy was deployed at the Barneo ice camp 2013. The resulting time series describes the evolution of snow height as a function of place and time between 2013-04-09 and 2013-06-13 in sample intervals of 1 hour. The Snow Depth Buoy consists of four independent sonar measurements representing the area (approx. 10 m**2) around the buoy. The buoy was installed on multi year ice. In addition to snow height, geographic position (GPS), barometric pressure, air temperature, and ice surface temperature were measured. Negative values of snow height occur if surface ablation continues into the sea ice. Thus, these measurements describe the position of the sea ice surface relative to the original snow-ice interface. Differences between single sensors indicate small-scale variability of the snow pack around the buoy. The data set has been processed, including the removal of obvious inconsistencies (missing values). Records without any snow height may still be used for sea ice drift analyses.

Snow height on sea ice and sea ice drift from autonomous measurements from buoy 2013S4, deployed during SIZONet 2013

Snow height was measured by the Snow Depth Buoy 2013S4, an autonomous platform, installed on land-fast sea ice off Barrow, Alaska during SIZONet 2013. The resulting time series describes the evolution of snow height as a function of place and time between 2013-04-09 and 2013-06-28 in sample intervals of 1 hour. The Snow Depth Buoy consists of four independent sonar measurements representing the area (approx. 10 m**2) around the buoy. The buoy was installed on land-fast sea ice. In addition to snow height, geographic position (GPS), barometric pressure, air temperature, and ice surface temperature were measured. Negative values of snow height occur if surface ablation continues into the sea ice. Thus, these measurements describe the position of the sea ice surface relative to the original snow-ice interface. Differences between single sensors indicate small-scale variability of the snow pack around the buoy. The data set has been processed, including the removal of obvious inconsistencies (missing values). Records without any snow height may still be used for sea ice drift analyses. Note: This data set contains only relative changes in snow height, because no initial readings of absolute snow height are available.

Snow height on sea ice and sea ice drift from autonomous measurements from buoy 2013S6, deployed during POLARSTERN cruise ANT-XXIX/6 (PS81)

Snow height was measured by the Snow Depth Buoy 2013S6, an autonomous platform, drifting on Antarctic sea ice, deployed during POLARSTERN cruise ANT-XXIX/6 (PS81). The resulting time series describes the evolution of snow height as a function of place and time between 2013-06-24 and 2013-09-27 in sample intervals of 1 hour. The Snow Depth Buoy consists of four independent sonar measurements representing the area (approx. 10 m**2) around the buoy. The buoy was installed on first year ice. In addition to snow height, geographic position (GPS), barometric pressure, air temperature, and ice surface temperature were measured. Negative values of snow height occur if surface ablation continues into the sea ice. Thus, these measurements describe the position of the sea ice surface relative to the original snow-ice interface. Differences between single sensors indicate small-scale variability of the snow pack around the buoy. The data set has been processed, including the removal of obvious inconsistencies (missing values). Records without any snow height may still be used for sea ice drift analyses.

Snow height on sea ice and sea ice drift from autonomous measurements from buoy 2013S8, deployed during POLARSTERN cruise ANT-XXIX/6 (PS81)

Snow height was measured by the Snow Depth Buoy 2013S8, an autonomous platform, drifting on Antarctic sea ice, deployed during POLARSTERN cruise ANT-XXIX/6 (PS81). The resulting time series describes the evolution of snow height as a function of place and time between 2013-07-09 and 2014-01-05 in sample intervals of 1 hour. The Snow Depth Buoy consists of four independent sonar measurements representing the area (approx. 10 m**2) around the buoy. The buoy was installed on first year ice. In addition to snow height, geographic position (GPS), barometric pressure, air temperature, and ice surface temperature were measured. Negative values of snow height occur if surface ablation continues into the sea ice. Thus, these measurements describe the position of the sea ice surface relative to the original snow-ice interface. Differences between single sensors indicate small-scale variability of the snow pack around the buoy. The data set has been processed, including the removal of obvious inconsistencies (missing values). Records without any snow height may still be used for sea ice drift analyses.