Research Sampling Techniques Quiz

Experimental & Quasi-Experimental Research

• This chapter covers experimental and quasi-experimental research methods.
• It details how to select samples and control for internal and external validity.

Learning Objectives

• How to select a sample for research validity.
• Controlling for internal and external validity.

Ways to Select a Sample

• A sample is the group of participants, treatments, and situations researched.
• Random sampling and stratified random sampling are ideal ways to select a sample that represents the population.
• Random sampling is often difficult, potentially impossible, due to challenges in contacting the entire population.

Random Sampling

• The purpose of random sampling is to ensure the sample represents the wider population, allowing inferences from the sample to be projected to the broader population.
• Random sampling helps ensures representativeness of the sample.
• Population is the entire group for research questions.

Example – Protein Supplementation

• Protein supplementation after competition maintains fat-free mass in female hockey players throughout the season.
• This observation can be expected in other women's U-Sport hockey programs.

Example – Random Sampling

• Data shows the typical error in DEXA, gym+field hockey (12 months), and field hockey (4 months).

Stratified Random Sampling

• Stratified random sampling first stratifies a population before random selection of the sample.
• Stratifying involves organizing into stable layers based on specific characteristics.
• Example: Studying physical fitness in grades 4-8 requires stratifying the population by age to create a relevant sample to gather normative data.

Voluntary Sample

• Researchers often rely on volunteers (recruited via flyers, media, online ads, or in-person contact).
• Volunteers are more interested in a study compared to a randomly selected sample.
• Findings from this method might not generalize to the wider population.

Convenience Sampling

• Convenience sampling selects participants readily available.
• Specific types of people might not be present in large numbers using random methods.
• Example: Canadian national team sport programs usually have 18-25 athletes, not a complete representation of the general population.

Snowball Sampling

• Researchers recruit from extremely specific groups (e.g., specific chronic diseases).
• The initial sample might be from convenience sampling.
• Current participants recruit other eligible individuals via word of mouth.
• Example: Examining effects of chemotherapy on cardiovascular function in breast cancer patients, enlisting patients during cancer survival meetings.

Ways to Assign Participants to Groups

• Selecting participants for control and treatment groups is done through random sampling and random assignment.
• Random assignment does not relate to random sampling. These are independent processes.

Random Assignment

• All true experimental designs require random assignment of groups within the sample.
• Random assignment criteria do not change based on methods used to gather the initial sample.

Random Assignment Procedures

• Each participant receives a unique ID (01-40 for 40 participants in the study) to be assigned to groups.
• Random assignment generators (MS Excel) are used to assign participants to groups.
• This process creates equivalent groups at the study baseline.

Random Matched Assignment

• First, match participants across groups based on a specific characteristic.
• The aim is to ensure each group has a similar mean age, VO2 Max, etc., to account for a selection bias.

Intact Groups

• Researchers sometimes must assign participants to existing groups (based on a condition or pre-existing groups).
• This method is not considered experimental.

Post Hoc Justifications

• Describes why a sample is considered representative of the population.
• This occurs after the event (experiment) has occurred.

Sources of Invalidity within Research Methods

• Internal validity: Determines if the experimental treatments created a difference in the instance.
• External validity: Assesses generalizability to other populations, settings, or treatments.

Internal Validity

• Ensuring internal validity involves controlling all independent variables.
• Controlling independent variables helps to minimize any rival explanations of the observed result in the dependent variable.
• However, gaining internal validity can make it more difficult to generalize (external validity) results to a wider population.

Example – Internal Validity

• Shows data points for the effects of yoga on plasma volume and cardiovascular performance in elite female hockey players.

Threats to Internal Validity

• History: External events unrelated to the treatment affecting the outcome.
• Maturation: Changes in participants over time (e.g., fatigue).
• Testing: Effects of repeated testing.
• Instrumentation: Changes in measure.
• Statistical regression: Selection based on extreme scores.
• Selection biases: Groups aren't formed randomly..
• Experimental mortality: Participants dropping out for reasons unrelated to the treatment.
• Selection-maturation interaction: Differences in characteristics between groups influencing maturation.
• Expectancies: Researchers' expectations influencing outcomes.

Threats to External Validity

• Reactive or interactive effects of testing: Participants modify their responses to subsequent treatments because the pre-test.
• Interaction of selection bias and experimental treatment: Treatment effects vary depending on participant characteristics.
• Reactive effects of experimental treatment: Treatments effective in constrained situations might not be in less constrained ones.
• Hawthorne effect: Participants' performance changes because of attention.
• Multiple-treatment interference: Previous treatments influence subsequent ones.

Controlling Threats to Internal Validity

• Randomization: Ensuring historical events, maturation, and statistical regression are evenly distributed across the groups.
• Placebo: Control group receives a false treatment.
• Blind setup: Participants don't know which treatment they are receiving.
• Double-blind setup: Neither participants nor researchers know which treatment is given.

Controlling for Internal Validity– Gold Standard

• Randomized, double-blind controlled studies are considered the gold standard in intervention-based research.
• The main advantage of using randomized, double-blind, controlled trials (DBRCT) is the ability to establish cause and effect relationships between variables.

Test your knowledge on various research sampling techniques through this engaging quiz. Explore concepts such as random sampling, stratified sampling, and convenience sampling, and how they influence research validity. Perfect for students and professionals alike!