Survey estimation and marginal effects in Stata

Stata is a general purpose software package that has become popular among various disciplines such as epidemiology, economics, or social sciences. Users like Stata for its scientific approach, its robustness and reliability, and the ease with which its functionality can be extended by user written programs. In this talk I will first give a brief overview of the functionality of Stata and then discuss two specific features: survey estimation and predictive margins/marginal effects. Most surveys are based on complex samples that contain multiple sampling stages, are stratified or clustered, and feature unequal selection probabilities. Standard estimators can produce misleading results in such samples unless the peculiarities of the sampling plan are taken into account. Stata offers survey statistics for complex samples for a wide variety of estimators and supports several variance estimation procedures such as linearization, jackknife, and balanced repeated replication (see Kreuter and Valliant, 2007, Stata Journal 7: 1-21). In the talk I will illustrate these features using applied examples and I will also show how user written commands can be adapted to support complex samples. Complex can also be the models we fit to our data, making it difficult to interpret them, especially in case of nonlinear or non-additive models (Mood, 2010, European Sociological Review 26: 67-82). Stata provides a number of highly useful commands to make results of such models accessible by computing and displaying predictive margins and marginal effects. In my talk I will discuss these commands provide various examples demonstrating their use.

Adaptation to Changes in the Coffee Value Chain and the Price of Coiffee among Coffee Producers in Two Villages in Kilimanjaro, Tanzania

Market liberalization in Tanzania has eroded the monopoly of the cooperative unions by allowing private coffee buyers (PCBs) to compete with them on equal footing. Similarly, farmers groups and primary societies are now allowed to sell coffee at auction. Thus, farmers have various options for selling their coffee. Similarly, the coffee industry has experienced large fluctuations in prices and stagnation in production. How do farmers react to these changes? Can and do farmers profit from different market conditions and sell to different traders at the lower end of the value chain, or do they remain with cooperatives or farmers groups? This study was conducted in Mruwia and Mshiri villages in Moshi Rural district. Whereas Mshiri village remains attached to the Kilimanjaro Native Cooperative Union (KNCU), Mruwia has detached from this organization and sells coffee independently. The sample (103) was randomly selected from the coffee farmers in the two villages. Data were collected through surveys, focus group discussions (FGDs), and socio-anthropological methods (participant-observation, biographies, and thematic interviews). Results indicate that the selection of whom to sell coffee depends largely on farmers’ dependence on coffee and prices, other benefits accrued, and whether the initial costs are covered by buyers. Additionally, most respondents did not sell coffee to PCBs. Thus, prices, the institutional infrastructure, and the structure of local communities were important when making decisions about how and with whom to trade.

The significant effect of endometriosis on physical, mental and social wellbeing: results from an international cross-sectional survey.

STUDY QUESTION To what extent do the management of endometriosis and the symptoms that remain after treatment affect the quality of life in women with the disease? SUMMARY ANSWER Many women with endometriosis had impaired quality of life and continued to suffer from endometriosis-associated symptoms even though their endometriosis has been managed in tertiary care centres. WHAT IS KNOWN ALREADY The existing literature indicates that quality of life and work productivity is reduced in women with endometriosis. However, most studies have small sample sizes, are treatment related or examine newly diagnosed patients only. STUDY DESIGN, SIZE, DURATION A cross-sectional questionnaire-based survey among 931 women with endometriosis treated in 12 tertiary care centres in 10 countries. PARTICIPANTS/MATERIALS, SETTING, METHODS Women diagnosed with endometriosis who had at least one contact related to endometriosis-associated symptoms during 2008 with a participating centre were enrolled into the study. The study investigated the effect of endometriosis on education, work and social wellbeing, endometriosis-associated symptoms and health-related quality of life, by using questions obtained from the World Endometriosis Research Foundation (WERF) GSWH instrument (designed and validated for the WERF Global Study on Women's Health) and the Short Form 36 version 2 (SF-36v2). MAIN RESULTS AND THE ROLE OF CHANCE Of 3216 women invited to participate in the study, 1450 (45%) provided informed consent and out of these, 931 (931/3216 = 29%) returned the questionnaires. Endometriosis had affected work in 51% of the women and affected relationships in 50% of the women at some time during their life. Dysmenorrhoea was reported by 59%, dyspareunia by 56% and chronic pelvic pain by 60% of women. Quality of life was decreased in all eight dimensions of the SF-36v2 compared with norm-based scores from a general US population (all P < 0.01). Multivariate regression analysis showed that number of co-morbidities, chronic pain and dyspareunia had an independent negative effect on both the physical and mental component of the SF-36v2. LIMITATIONS, REASONS FOR CAUTION The fact that women were enrolled in tertiary care centres could lead to a possible over-representation of women with moderate-to-severe endometriosis, because the participating centres typically treat more complex and referred cases of endometriosis. The response rate was relatively low. Since there was no Institute Review Board approval to do a non-responder investigation on basic characteristics, some uncertainty remains regarding the representativeness of the investigated population. WIDER IMPLICATIONS OF THE FINDINGS This international multicentre survey represents a large group of women with endometriosis, in all phases of the disease, which increases the generalizability of the data. Women still suffer from frequent symptoms, despite tertiary care management, in particular chronic pain and dyspareunia. As a result their quality of life is significantly decreased. A patient-centred approach with extensive collaboration across disciplines, such as pain specialists, psychologists, sexologists and social workers, may be a valuable strategy to improve the long-term care of women with endometriosis. STUDY FUNDING/COMPETING INTEREST(S) The WERF EndoCost study is funded by the World Endometriosis Research Foundation (WERF) through grants received from Bayer Schering Pharma AG, Takeda Italia Farmaceutici SpA, Pfizer Ltd and the European Society of Human Reproduction and Embryology. The sponsors did not have a role in the design and conduct of the study; collection, management, analysis and interpretation of the data; and preparation, review or approval of the manuscript. L.H. is the chief executive and T.D. was a board member of WERF at the time of funding. T.D. holds the Merck-Serono Chair in Reproductive Medicine and Surgery, and the Ferring Chair in Reproductive Medicine at the Katholieke Universiteit Leuven in Belgium and has served as consultant/research collaborator for Merck-Serono, Schering-Plough, Astellas and Arresto.

Does World Natural Heritage status trigger sustainable regional development efforts?

Today, there are over 200 World Natural Heritage (WNH) sites. Although the original aim of the World Heritage (WH) Convention was to spark off concerted international efforts to preserve sites of outstanding and universal value, today a multitude of expectations rests on WNH sites in terms of conservation, tourism, management and regional development. This paper identifies the effects of WNH status on sustainable regional development and the driving factors behind these effects. The results are based on a global survey of WNH sites and qualitative interviews with key WNH personnel. The paper shows that WNH status can be an important trigger for sustainable regional development, but its effectiveness depends on a number of intricately interwoven ‘soft’ success factors. Clearer policies and management guidelines, as envisaged by UNESCO, are crucial to achieving a balance between conservation and development.

Combined effect of soil bund with biological soil and water conservation measures in the northwestern Ethiopian highlands

Excessive runoff and soil erosion in the upper Blue Nile Basin poses a threat that has attracted the attention of the Ethiopian government because of the serious on-site effects in addition to downstream effects, such as the siltation of water harvesting structures and reservoirs. The objective of the study was to evaluate and recommend effective biophysical soil and water conservation measure(s) in the Debre Mewi watershed, about 30 km south of the Lake Tana. Six conservation measures were evaluated for their effects on runoff, soil loss, and forage yield using runoff plots. There was a significant difference between treatments for both runoff and soil loss. The four-year average annual soil loss in the different plots ranged from 26 to 71 t ha−1, and total runoff ranged from 180 to 302 mm, while annual rainfall varied between 854 mm in 2008 and 1247 mm in 2011. Soil bund combined with elephant grass had the lowest runoff and soil loss as compared to the other treatments, whereas the untreated control plot had the highest for both parameters. As an additional benefit, 2.8 and 0.7 t ha−1 year−1 of dried forage was obtained from elephant and local grasses, respectively. Furthermore, it was found that soil bund combined with Tephrosia increased soil organic matter by 13% compared to the control plot. Soil bund efficiency was significantly enhanced by combining them with biological measures and improved farmers’ perception of soil and water conservation measures.

ADDPLOT: Stata module to add twoway plot objects to an existing twoway graph

addplot adds twoway plot objects to an existing twoway graph. This is useful if you want to add additional objects such as titles or extra data points to a twoway graph after it has been created. Most of what addplot can do, can also be done by rerunning the original graph command including additional options or plot statements. addplot, however, might be useful if you have to modify a graph for which you cannot rerun the original command, for example, because you only have the graph file but not the data that were used to create the graph. Furthermore, addplot can do certain things that would be difficult to achieve in a single graph command (e.g. customizing individual subgraphs within a by-graph). addplot also provides a substitute for some of the functionality of the graph editor.

Vultures vs livestock: conservation relationships in an emerging conflict between humans and wildlife

Human–wildlife conflict is emerging as an important topic in conservation. Carnivores and birds of prey are responsible for most conflicts with livestock and game but since the mid 1990s a new conflict is emerging in south-west Europe: the presumed killing of livestock by griffon vultures Gyps fulvus. Lack of scientific data and magnification of the problem by the media are increasing alarm amongst the public, and political pressures to implement management decisions have not been based on scientific evidence. We compiled information on 1,793 complaints about attacks by griffon vultures on livestock, lodged with Spanish authorities from 1996 to 2010. Spain is home to the majority (95%) of griffon vultures and other scavengers in the European Union. Most of the cases occurred in areas of high livestock density, affected principally sheep (49%) and cows (31%), and were associated with spring birthing times (April–June). On average 69% of the complaints made annually were rejected because of a lack of evidence about whether the animal was alive before being eaten. The total economic cost of compensation was EUR 278,590 from 2004 to 2010. We discuss possible ways to mitigate this emerging human–wildlife conflict. These need to include the participation of livestock farmers, authorities, scientists and conservation groups.

Bright moonlight triggers natal dispersal departures

Upon leaving their natal area, dispersers are confronted with unknown terrains. Species-specific perceptual ranges (i.e. the maximum distance from which an individual can perceive landscape features) play a crucial role in spatial movement decisions during such wanderings. In nocturnal animals that rely on vision, perceptual range is dramatically enhanced during moonlight, compared to moonless conditions. This increase of the perceptual range is an overlooked element that may be responsible for the successful crossing of unfamiliar areas during dispersal. The information gathered from 143 radio-tagged eagle owl Bubo bubo juveniles in Spain, Finland and Switzerland shows that, although the decision to initiate dispersal is mainly an endogenous phenomenon determined by the attainment of a given age (∼6 months), dispersers leave their birthplace primarily under the best light conditions at night, i.e. when most of the lunar disc is illuminated. This sheds new light into the mechanisms that may trigger dispersal from parental territory.

ROBREG: Stata module providing robust regression estimators

robreg provides a number of robust estimators for linear regression models. Among them are the high breakdown-point and high efficiency MM-estimator, the Huber and bisquare M-estimator, and the S-estimator, each supporting classic or robust standard errors. Furthermore, basic versions of the LMS/LQS (least median of squares) and LTS (least trimmed squares) estimators are provided. Note that the moremata package, also available from SSC, is required.

Three essays on the concept of trust and its foundations

The present dissertation focuses on trust and comprises three empirical essays on the concept itself and its foundations. All three essays investigate trust as an expectation and rely on selfreport measures of trust. Whereas the first two chapters investigate social trust, the third chapter investigates political trust. Essentially, there are three related important debates to which the following chapters contribute. A first debate discusses problems with current selfreport measures. Scholars recently started to question whether standard trust questions really measure the same across countries and languages. Chapter 1 engages in this debate. Using data from Switzerland it studies whether different trust questions measure the same latent trust constructs across individuals belonging to three different culturallinguistic regions. The second debate concerns the socalled forms or dimensions of trust. Recently, scholars started investigating whether trust is a onedimensional construct, i.e. whether an individual's trust judgment differs for categories of trustees such as strangers, neighbors, family members and friends or not. Relying on confirmatory factor analysis Chapter 2 investigates whether individuals really do make a difference between different trustee categories and to what extent these judgments can be summarized into higherorder latent trust constructs. The third debate is concerned with causes of differences in trust across humans. Chapter 3 focuses on the role of laterlife experiences, more precisely victimization experiences and investigates their causal relationship with generalized social trust. Chapter 4 focuses on the impact of direct democratic institutions on the trust relationship between citizens and political authorities.

Constitutionality: Conditions for Crafting Local Ownership of Institution-Building Processes

This article presents constitutionality as a new approach for analyzing bottom-up institution-building processes emphasizing local perceptions and local agency in common pool resource management. Using four case studies—fisheries in Zambia; pasture and forestry in Mali; fisheries in Indonesia; forestry in Bolivia—this approach analyzes examples of local institution building differing from top-down imposed participation. Our analysis highlights six components of constitutionality: emic perceptions of the need for new institutions, participatory processes of negotiation, preexisting institutions as a basis for institution building, outside catalyzing agents, recognition of local knowledge, and higher level acknowledgment of the new institutions.

PSHARE: Stata module to compute and graph percentile shares

-pshare- computes and graphs percentile shares from individual level data. Percentile shares are often used in inequality research to study the distribution of income or wealth. They are defined as differences between Lorenz ordinates of the outcome variable. Technically, the observations are sorted in increasing order of the outcome variable and the specified percentiles are computed from the running sum of the outcomes. Percentile shares are then computed as differences between percentiles, divided by total outcome. pshare requires moremata to be installed on the system; see ssc describe moremata.

Technical Efficiency in Public and Private Sectors in India: Evidence from the Post-Reform Years

While India's state-owned enterprises are widely believed to be inefficient, there is a dearth of studies that document such inefficiency on any rigorous basis. Yet, since improvement in firm efficiency is one of the basic objectives of privatization, it is important to assess whether efficiency is indeed lower in the public sector than in the private sector. This paper compares the performance of state-owned enterprises with those of private sector firms in respect of technical efficiency. The comparison is made in eight different sectors over the period 1991-92 to 1998-99. We measure technical efficiency using the method of Data Envelopment Analysis. Judging by the average levels of technical efficiency, no conclusive evidence of superior performance on the part of the private sector is found.

LORENZ: Stata module to estimate and display Lorenz curves and concentration curves

Lorenz estimates Lorenz and concentration curves from individual-level data and, optionally, displays the results in a graph. Relative as well as generalized, absolute, unnormalized, or custom-normalized Lorenz or concentration curves are supported, and tools for computing contrasts between different subpopulations or outcome variables are provided. Variance estimation for complex samples is fully supported.