Research Design - Semantic 1

Definitions, concepts, facts

Instrumentation effects

Refer to the potential biases or changes in research results caused by the tools or measures used to collect data. These effects may alter participant responses or influence the validity of the findings. Internal validity threat.

Demoralization bias

Occurs when participants lose motivation or enthusiasm, leading to altered responses or behaviors that may negatively impact study results. Common with control groups who believe they are being deprived of beneficial treatment.

CONSORT table

A table that summarizes the flow of participants through a clinical study, detailing how many were assessed for eligibility, allocated to each group, and those who completed or were lost to follow-up. It is used to enhance transparency and report findings in accordance with the CONSORT statement.

CONSORT acronym

Consolidated Standards of Reporting Trials, system that standardizes how data from Randomized Clinical Trials is reported.

External validity

The extent to which study findings can be generalized to, or have relevance for settings, people, times, and measures other than the ones studied. It assesses how well research outcomes apply outside the specific conditions of the study.

Effectiveness research

A type of research that evaluates how well a specific intervention works in real-world settings, considering factors such as cost, quality of life, and feasibility. It attempts to maximize external validity.

Efficacy research

A type of research that tests how well an intervention works under controlled conditions, often in clinical trials, to determine its potential effectiveness. It attempts to maximize internal validity.

Retrospective case-control design

A type of observational study that compares individuals with a specific condition (cases) to those without it (controls) by looking back at their exposure to potential risk factors.

Demand characteristics

are cues that influence participants' behavior in a study, often leading them to respond in a way they believe aligns with the researcher's expectations.

Prospective case-control design

A type of observational study that compares participants with a specific condition (cases) to those without (controls) by following them forward in time to assess exposure to potential risk factors.

Stratified random sampling

A sampling method that divides the population into subgroups (strata) and randomly selects samples from each stratum to ensure representation across key characteristics. Form of probability sampling.

Mono-method bias

refers to the potential distortions in research findings caused by relying on a single method of data collection, which threatens construct validity.

Historical control group design

A quasi-experimental study design that uses existing data from a group received a treatment or intervention in the past as a comparator for a current group receiving the same treatment, often to assess effectiveness or safety.

Treatment fidelity

The extent to which an intervention is delivered as intended, ensuring consistency in implementation across different settings or participants. Avoids modifying or omitting material.

Splitting the file

is a process used in qualitative research to separate data into different segments for analysis, ensuring distinct themes or categories are identified. Common for dividing a dataset across different levels of a confounding variable.

Moderator

A variable that affects the strength or direction of the relationship between an independent and a dependent variable, potentially altering the outcomes of the research study.

Rosenthal effect

The phenomenon where researchers' expectations influence their observations or participants' performance, potentially skewing study results.

Bias

is systematic error variance in the dataset that limits generalizability of results.

Hypothesis guessing

is a cognitive bias where participants attempt to deduce the purpose of a study, potentially influencing their behavior and the study's results.

Evaluation apprehension

is a form of bias where participants worry about how their responses will be evaluated, which can affect their behavior and impact study outcomes.

Case-control design

is a type of observational study that compares individuals with a specific condition or outcome (cases) to those without it (controls), often assessing exposure to potential risk factors. Condition is the dependent variable, risk factors are the independent variables.

Cluster-randomized design

is a type of quasi-experimental study design where groups or clusters, rather than individuals, are randomly assigned to different treatment conditions. This approach is often used to evaluate interventions in communities or groups.

Cohort design

is an observational study design that follows a group of individuals (cohort) over time to assess the effects of various exposures on specific outcomes, allowing for the examination of causality.

Intact groups design

is a type of quasi-experimental design where existing groups are used for comparison rather than randomly assigning participants. This design is often applied in educational, correctional, institutional, and other social contexts.

Mediator

is a variable that explains the mechanism through which an independent variable influences a dependent variable, often helping to clarify the relationship between the two.

Funnel plot

is a graphical tool used to assess publication bias in meta-analyses by plotting effect sizes against their standard errors, typically resulting in a funnel shape when there is no bias.

Mono-operational bias

is the result of limited operational definitions for a concept, leading to a skewed understanding and measurement of that concept's effect in research, threatening construct validity.

Expectancy effects

are changes in participants' behavior or performance due to their expectations about the treatment or outcome, potentially impacting study results. Most common example is the placebo effect.

Cross-sectional case-control design

is an observational study design where participants are measured at one point in time. Results of measures are coded as dependent variables, and the independent variable is considered to be a hypothesized cause of those scores.

Fail safe n

is an estimated minimum number of unpublished/unincluded studies with negative findings that would be needed to render the mean effect size of a meta-analysis statistically insignificant.