Pengantar Penelitian Eksperimen

Introduction to Experimental Research

Experimental research is a form of quantitative research used to test whether a treatment or manipulation causes a change in an outcome. Researchers compare outcomes across groups receiving different treatments, or compare a treated group with one that receives no treatment.

Experimental Research Compared with Survey Research

• Experimental research tests cause and effect. The researcher manipulates an independent variable and measures or observes its effect on a dependent variable.
• Survey research describes or measures attitudes, opinions, or other existing characteristics, commonly using questionnaires or self-reports. It does not involve a treatment imposed by the researcher.
• Researchers often refer to people in experiments as participants and people in surveys as respondents.
• For example, a study of whether different types of service recovery affect airline customers’ word of mouth treats service recovery as the manipulated variable and word of mouth as the outcome.

Three Defining Features of Experiments

1. Manipulation: The researcher deliberately provides a treatment or varies a condition.
2. Random assignment: In a true experiment, participants have an equal chance of being assigned to different treatment groups.
3. Control: Researchers try to hold other relevant influences constant so that observed effects can more confidently be attributed to the manipulation.

True Experiments and Quasi-Experiments

A true experiment uses random assignment to treatment groups. A quasi-experiment does not randomly assign participants; instead, the researcher assigns participants to groups or uses groups that are already arranged.

A quasi-experiment may be chosen when interaction between participants could contaminate the results. For instance, comparing two teaching methods among students at the same school could allow students to share information. The video illustrates assigning one method to School A and another method to School B instead.

Key Variables

• Independent variable: The factor manipulated by the researcher and treated as the potential cause.
• Dependent variable: The outcome measured or observed and treated as the possible effect.
• Extraneous variables: Other factors that could influence the outcome but are not the focus of the study. Researchers should control these or keep them as constant as possible.
• Confounding variables: Extraneous factors that were not adequately controlled and may offer an alternative explanation for the results, weakening the study’s internal validity.

For example, a fertilizer experiment should account for other influences on plant growth, such as sunlight, water, temperature, humidity, seed type, and pests. Similarly, a study of television content and children’s aggression should consider other relevant influences, such as personality and prior bullying experiences.

Treatment Levels and Study Groups

• An independent variable should have two or more levels—different versions or conditions of the treatment. For example, a study might compare violent, humorous, and dramatic television content.
• An experimental group receives a manipulation.
• A control group receives no treatment and can serve as a comparison. Not every experiment must have a control group.

Forms of Manipulation

The video describes three common ways to manipulate conditions:

• Situational manipulation: Place participants in a designed situation or scenario.
• Task manipulation: Ask participants to perform a task, such as solving a puzzle.
• Instructional manipulation: Give participants instructions intended to produce a particular response or state, such as imagining a scenario.

Measuring Outcomes

Dependent variables can be assessed through more than questionnaires or self-report scales:

• Self-report measures: Participants describe their own experiences or responses.
• Behavioral measures: Examples include reaction time, movement speed, or other observable actions.
• Physiological measures: Examples include skin response, heart rate, ECG, or MRI.

The presenter notes that psychological research can benefit from measures of observable behavior and physiology rather than relying exclusively on self-report.

Preliminary Trials and Manipulation Checks

• Pilot studies are small-scale trials conducted before the main experiment. They can test stimuli, procedures, and instructions, and identify whether the setup seems natural or presents practical obstacles.
• A manipulation check tests whether participants experienced or understood the intended treatment. It is typically conducted after the outcome measurement. If participants in a group do not perceive the manipulation as intended, their data may be excluded from further analysis.

Research Ethics

• Experimental procedures must not harm participants or create unjustified negative effects.
• Participants should receive information about the procedure and their rights so they can choose voluntarily whether to take part. This is informed consent.
• In some studies, researchers use deception—withholding or misrepresenting the purpose or procedures to avoid influencing participants’ behavior.
• When deception is used, researchers should provide a debriefing afterward and explain the actual purpose and procedures. They may also ask participants to keep details confidential temporarily so later participants are not influenced.

Speakers and Sources Mentioned

• Speaker/presenter: Resekiani Mas Bakar
• Cited source: John W. Creswell, Research Design: Qualitative, Quantitative, and Mixed Methods Approaches (2014)
• Researcher mentioned: Roy Baumeister, in connection with the critique of relying only on self-report in psychology