JOSÉ SUPO - DISEÑOS DE INVESTIGACIÓN

Diseño de Investigación: Conceptos Clave

Introducción al Tema

• José S. inicia la transmisión en vivo, solicitando a los participantes que confirmen la recepción del audio y compartan su ubicación.

• Se establece que el enfoque de la sesión será sobre "diseños de investigación", sin entrar en tipos o niveles de investigación.

Definición y Importancia del Diseño

• Un diseño se define como una estrategia específica para desarrollar una idea de investigación; no todas las ideas generan intenciones investigativas.

• La necesidad de tener una línea de investigación es crucial para que surjan ideas que inviten a la acción y desarrollo de estudios.

Comparación con Estrategias Deportivas

• Se compara el diseño de investigación con un traje adecuado para practicar karate; primero se necesita una idea antes de elegir un diseño.

• Así como un equipo deportivo adapta su estrategia según el rival, los investigadores deben ajustar sus diseños según las ideas surgidas.

Tipos y Taxonomía de Diseños

• Existen infinitos diseños posibles, pero se pueden clasificar en grupos basados en su origen dentro del campo del conocimiento.

• Los diseños se agrupan en epidemiológicos, experimentales, comunitarios/ecológicos y validación de instrumentos, cada uno con características únicas.

Aplicaciones Prácticas

• La mayoría de investigaciones académicas involucran sujetos humanos; los diseños epidemiológicos estudian personas mientras que otros pueden estudiar objetos o eventos.

• Los diseños comunitarios analizan conjuntos humanos, mientras que los experimentales evalúan eventos relacionados con individuos.

Características Esenciales de Diseños Epidemiológicos

Enfoque Principal

• Los diseños epidemiológicos se centran en morbilidad (enfermedad) y mortalidad (muerte), siendo fundamentales en ciencias de la salud.

Aplicabilidad Más Allá del Campo Médico

• Aunque originados en ciencias médicas, estos diseños pueden aplicarse a otras áreas como educación o ciencias sociales bajo ciertas condiciones éticas.

Clasificación General

• Se dividen clásicamente en dos grupos: descriptivos (buscan validez interna sobre un grupo específico) y analíticos (buscan validez externa extrapolable).

Diferencias entre Estudios Descriptivos y Analíticos

Validez Interna vs. Externa

• Los estudios descriptivos buscan conclusiones válidas solo sobre el grupo estudiado; por otro lado, los analíticos permiten extrapolar resultados a poblaciones más amplias.

Ejemplos Prácticos

• Un estudio sobre fumadores puede concluir que fumar es un riesgo para cáncer pulmonar no solo localmente sino globalmente si logra establecer relaciones entre variables.

Diseños Observacionales vs. Experimentales

Estudio Observacional

• Incluye casos controles (retrospectivo/transversal), donde se observa sin manipulación directa; útil para evaluar factores riesgosos aunque menos robusto comparado con estudios experimentales.

Estudio Experimental

• Implica manipulación controlada para observar efectos específicos; incluye ensayos clínicos divididos en fases enfocadas tanto a individuos como poblaciones.

Elementos Fundamentales del Diseño Experimental

Manipulación Controlada

• La manipulación es esencial; implica intervenir deliberadamente para observar cambios causales. Sin esta intervención intencionada no hay experimento válido.

Consideraciones Éticas

• Al realizar experimentos con seres humanos hay limitaciones éticas significativas que deben ser consideradas para proteger la integridad física y mental.

Understanding Experimental Designs in Research

Treatment and Control in Experiments

• The term "treatment" refers to the intervention or active principle applied in experimental designs, which can involve plants, animals, or other subjects.

• Methodological control is essential in experiments; it involves eliminating confounding variables and applying statistical controls like multivariate analysis.

• To establish a causal relationship, criteria such as temporal precedence must be met; for example, smoking precedes lung cancer.

• Control methods include methodological (eliminating confounding variables) and statistical (multivariate analysis), crucial for defining the study population and sample frame.

Characteristics of True Experiments

• A true experiment requires randomization of units of study into treatment groups to ensure valid comparisons.

• The concept of "unit of study" is broad; it can refer to individuals, objects, or events rather than just human subjects.

• Distinction between unit of study (e.g., a pregnant woman) and experimental unit (e.g., childbirth event).

Randomization and Its Importance

• Randomly dividing participants into treatment and control groups helps eliminate bias related to pre-existing differences among subjects.

• Without randomization, results may reflect inherent differences rather than the effect of the treatment itself.

Measurement Before and After Intervention

• Baseline measurements are critical; they allow comparison before and after an intervention to assess its impact effectively.

• Internal control compares initial measures with final outcomes post-intervention; this is also referred to as autocontrol.

Statistical Analysis in Experimental Design

• The independent variable should be categorical while the dependent variable should be numerical for effective measurement post-intervention.

• Statistical tests like Student's t-test are used to compare baseline measures with final outcomes across different groups.

Challenges in Experimental Design

Addressing Non-Randomized Studies

• In cases where randomization isn't possible due to ethical concerns (e.g., withholding treatment), researchers must acknowledge limitations when interpreting results.

Quasi-experiments: Definition and Limitations

• Quasi-experiments lack full randomization but still involve manipulation. They often have a single group with pre-and post-measurements without a control group.

Pre-experimental Designs: An Overview

• Pre-experiments do not have baseline measures but may still involve manipulation. They are generally less reliable than true experiments due to their design flaws.

Community or Ecological Designs

Characteristics of Community-Based Research

• Community designs analyze aggregated data from populations rather than individual-level data, focusing on broader health indicators like maternal mortality rates.

Retrospective Nature of Some Studies

• Many community studies are retrospective, relying on historical data rather than prospective collection methods that could mitigate biases.

Maternal Mortality Study Design

Understanding Maternal Mortality Rates

• The discussion begins with the concept of maternal mortality rates, emphasizing that these rates are calculated based on the number of women who died during pregnancy, childbirth, or postpartum compared to live births within a year.

Challenges in Prospective Studies

• A prospective study for maternal mortality is theoretically outlined but deemed impractical since it would require tracking pregnant women throughout their pregnancies and ensuring they receive proper health care to prevent deaths.

• The speaker highlights that identifying pregnant women and monitoring them over a year poses significant logistical challenges, making such studies difficult to implement effectively.

Retrospective Studies as an Alternative

• Due to the impracticality of prospective studies, maternal mortality research typically relies on retrospective designs using previously recorded data from health registries and statistics. This approach allows researchers to analyze historical data rather than following individuals prospectively.

Community-Based Research Designs

Characteristics of Social Sciences Research

• In social sciences, researchers often deal with subjective variables like quality of life and job satisfaction, contrasting with objective measures used in health sciences (e.g., weight or temperature). This distinction is crucial for understanding how different fields approach research methodologies.

Subjective vs Objective Variables

• The speaker emphasizes that while health sciences focus on measurable physical attributes, social sciences frequently analyze subjective constructs which can lead to categorical variables rather than numerical ones. This difference shapes how research findings are interpreted across disciplines.

Exploratory vs Analytical Studies

Types of Exploratory Research Designs

• Exploratory studies include ethnographic designs and focus groups aimed at understanding community behaviors without statistical analysis; they rely heavily on qualitative observations and interviews instead. Examples include studying lifestyle choices in specific communities like those living on floating islands in Lake Titicaca.

Ethnographic Studies Explained

• An ethnographic study involves immersive observation where researchers live among subjects for extended periods to gather insights into their behaviors and attitudes towards health issues without employing statistical methods for analysis.

Analytical Community Studies

Distinction Between With and Without Intervention

• Analytical studies can be categorized into those involving interventions (like public health programs) versus observational studies without interventions; both aim to assess population-level outcomes but differ in methodology and application context. Examples include evaluating government programs addressing endemic diseases like dengue fever or anemia among children in Peru.

Importance of Population Parameters

• The focus shifts toward analyzing group outcomes rather than individual metrics; this perspective helps determine the effectiveness of public health initiatives by examining overall population statistics rather than isolated cases.

Validation of Instruments in Behavioral Sciences

Defining Constructs for Measurement

• Validating instruments requires clear definitions of constructs being measured; ambiguity can lead to varied interpretations among respondents, complicating data collection efforts significantly when assessing subjective phenomena like political beliefs or consumer behavior.

Steps in Instrument Creation

• Creating measurement tools involves defining constructs clearly before developing items/questions; validation processes may involve expert reviews (validation by judges) or empirical testing depending on the clarity of the construct being assessed.

This structured markdown file captures key discussions from the transcript while providing timestamps for easy reference back to specific points made during the talk.

Reproducibility and Repeatability in Measurements

Understanding Key Concepts

• Repeatability is defined as the ability of a single evaluator to obtain consistent results when measuring the same individual multiple times with the same instrument.

• Reproducibility, on the other hand, involves multiple evaluators using the same instrument to measure the same individual and achieving consistent results. This highlights a concept of stability across different evaluators.

• The failure of repeatability may occur due to a lack of procedural manuals for applying instruments, while reproducibility can fail if evaluators are not adequately trained or prepared for measurements.

Statistical Analysis

• Both repeatability and reproducibility are assessed through statistical measures such as variance analysis for numerical variables, while categorical variables utilize different methods like Kappa statistics.

• Research designs, particularly R&R (reliability and reproducibility), are essential frameworks that guide these evaluations in research contexts.

Criterion Validity

Types of Criterion Validity

• Concurrent validity refers to comparing measured results against an external reference at the same time as screening tests.

• If immediate measurement isn't possible, researchers may use predictive validity, which assesses how well a test predicts outcomes based on an external reference.

Diagnostic Performance

• The diagnostic performance of an instrument can be evaluated using metrics like R-squared for numerical variables or ROC curves for categorical variables.

• Determining cut-off points in measurements often relies on criteria-based methods or probabilistic approaches such as ROC analysis.

Adaptation and Cultural Context

When to Adapt Instruments

• Instrument adaptation occurs when existing tools do not meet specific metric properties or require translation from one language to another.

• Effective translation requires understanding contextual meanings rather than literal translations; this ensures that terms resonate appropriately within cultural contexts.

Challenges in Translation

• Translating instruments necessitates expertise not only in language but also in the subject matter being assessed. This dual knowledge is crucial for accurate interpretation and application.

Updating Instruments Post-Pandemic

Impact of Changing Contextual Factors

• The concept of self-medication has evolved significantly during the pandemic, affecting how previously constructed instruments assess this behavior.

• Instruments developed before 2020 may need reevaluation due to shifts in public health practices and behaviors influenced by pandemic conditions.

Cultural Sensitivity in Adaptation

• Adapting instruments culturally involves recognizing differences in term interpretations across communities, ensuring that questions resonate with target populations effectively.

Research Design Considerations

Importance of Research Designs

• Understanding various research designs is critical as they provide structured strategies for addressing research questions. These include epidemiological, experimental, community-based designs among others.