Lec 5 - Understanding Quantitative Research Designs and Review Studies

Experimental Design

• Contains interventions, controlled groups, and are usually randomized.

  • Randomized Controlled Trials (RCT)

  • Quasi-experiments

Non-Experimental - Exploratory

• Observational study

• Finding patterns or possible relationships

Non-Experimental - Descriptive

• Observational study

• Simply describes what is happening

True Experimental

• Strictly follows the scientific method

  • Uses random assignments, which eliminates bias and increases the reliability of results

• The main purpose is to test cause-and-effect relationships

Manipulation

• Essential Features of Experimental Research

•  Researcher actively changes or introduces intervention to see the outcome

• When the independent variable is manipulated, it is said to be active

• Any difference between groups is considered causally related to group membership

  • The differences in the outcome between the experimental and controlled group is assumed to be caused by the treatment applied in the experimental group

Control

• Essential Features of Experimental Research

•  Having a controlled group that does not receive the new intervention

• Control condition is present to establish the null effect

  • Similar group, similar exposure, different treatment

  • Establishes a baseline to show what happens when there is no treatment

  • Blinding

Randomization

• Essential Features of Experimental Research

•  Randomly assigning participants to treatment or control groups, ensuring any differences between them are due to the intervention, not by chance or preference

• Group equality 

  • At the start, groups are expected to have equal/balanced characteristics

Randomized Controlled Trial

• Gold standard of experimental designs

  • Ensures groups are similar at the start of the study

Pretest-Posttest Control Group Design

• RCT Designs

• Most common and basic RCT design

• Commonly used in research studies, particularly in psychology and education

• Steps:

  • Randomly assigned participants

  • Both go through pretest

  • Experiment group receives intervention

  • Both go through posttest 

• The difference in the outcome between the groups can be attributed as the treatment’s effect

Two-group Pretest-Posttest Design

• RCT Designs

• Steps:

  • Randomly assigned participants

  • Both go through pretest

  • Experiment group receives intervention

  • Both go through posttest 

• The difference in the outcome between the groups can be attributed as the treatment’s effect

• There is no distinct specification to the control group

  • Both may be treated as experimental or comparison groups

Multi-group Pretest-Posttest Design

• RCT Designs

• Steps:

  • Participants are divided into multiple groups, for different conditions or treatment

  • Pretest is conducted for all groups

  • Each group receives a different intervention

  • Posttest is conducted for all groups

• The differences among the outcomes can be attributed to the effects of the treatment

• Allows researchers to compare effects of multiple interventions or conditions simultaneously

Quasi-Experimental

• Shares some characteristics with experimental designs, but lacks random assignments of participants

• Allows for the limitations of natural settings where randomization is difficult, impractical, or unethical

Systematic Review

• Synthesizes results of several studies that have sought to answer similar research questions

• Combines findings of multiple studies to provide a truer picture and whole results

• Most common among review designs

• Methodologically, strongest review design

• Features:

  • Precise review questions

  • Explicit selection criteria

  • Clear methods for literature search, quality assessment, data extraction, and data analysis

• Steps are well-defined and methods are explicitly reported for reproducibility

Meta-analysis

• A systematic review that involves quantitative synthesis of results

• Uses statistical pooling of findings across multiple studies to generate a single numerical summary

• Steps:

  • Gather all available evidence

  • Synthesize

  • It will come up as if you have conducted a bigger study 

Forest Plots

• Visual summary of whether trial results can be suitably combined into one numerical outcome

• Meta-analysis always have forest plots

• Illustrates the synthesized effects of the trials that are included in the systematic review.

Scoping Review

• Preliminary assessment of the potential size and scope of available research literature

• Identifies the nature and extent of research evidence

• Rigorous and systematic

• Identifies the knowledge gaps, allowing researcher to see what gaps exist and what directions need to be pursued

Narrative Review

• Involves some process for identifying and selecting materials for inclusion, for synthesizing in textual/tabular/graphical form, and for making some analysis of their contribution or value

• NOT considered rigorous

• Provides an overview or contextualization of the knowledge area

• Writers already have a comprehensive knowledge of the area

Exploratory Design

• Used to find the relationship between variables and outcomes

  • Generates hypotheses about certain phenomena

  • To clarify concepts

• Any difference between groups is merely an association or correlation, but does not establish cause and effect relationships

• Aims to better understand groups

  • Not much is known or builds on the existing knowledge of the topic

• Does NOT manipulate variables

• Answers: how and why

Descriptive Design

• Used to describe populations

  • Describe characteristics/important features of a population

  • Illustrate trends/patterns over time

  • Document conditions and events

• Aims to better understand groups

  • Not much is known about the topic

• Does NOT manipulate variables

• Answers: what, where, who, when, how much, and how many

Retrospective Exploratory Research

• Data has been collected previously, often for a different purpose

• Relationship between two or more variables is tested

• Easier to implement, but less control over the quality of data

Prospective Exploratory Research

• Data is collected as it comes

• Able to determine the relationship between variables as they occur

• Stronger in methodological rigor

• More control on how data is collected, specific variables, and design the data collection protocol

Longitudinal

• Following participants over a period of time and see changes through different data sets collected per time frame

• Begin with a baseline assessment of outcomes and follow up

• Seeing changes helps identify factors that could influence changes and propose relationships between variables or outcomes

• May be affected by:

  • Dropout or attrition, losing participants overtime

  • Testing effects, repeated testing means participants are able to learn the components and improve

  • Congounders with time, exposed to other events beyond control that may impact outcomes

    • Natural Recovery

    • Disease Progression

Cross-Sectional

• Using one set of data only as it is collected at one point in time

• May be affected by:

  • Selection, difficulty in recruiting similar participants, not being able to control the characteristics of the participants

  • Cohort effects, effects of recruitment of participants in the same age group

    • Similar experiences

    • Similar factors unique to the age group

Cohort Studies

• Follows one or more cohorts of individuals over a period of time

• Prospective or retrospective investigation

• Determines the cohort’s status with respect to an outcome of interest, and their exposure to risk factors

Developmental Research

• Investigation of patterns of growth and change over time

• May chronicle the natural history of a disease or disability

Normative Research

• Establishes normal values for specific variables

• Serves as guidelines for diagnosis and treatment planning

• Useful when we want to make sense of our patients in relation to what is expected of a healthy individual with similar conditions

Case Report or Case Series

• Description of one or several patients or communities to document unusual conditions or the effect of innovative interventions

• Difference lies in number of cases

• Provides information about feasibility of interventions, describes features of rare conditions, and help generate hypotheses for future research

• Used when something uncommon is encountered or when new approaches to care are applied

Historical Research

• Reconstructs the past to generate questions or suggest relationships of historical interest to inform contemporary perspectives

Qualitative Studies Research

• Also called naturalistic research

• Observations or interviews exploring the meaning of human experience as it occurs naturally

• Generates hypotheses about human behavior

Mixed Methods Research

• A combination of quantitative and qualitative research methods

Descriptive Surver Research

• Conducted through the use of self-report questionnaires

• Used in market research, where respondents are asked about their opinions on products or services