Research Design: Blueprint for Investigation

• Research design is a detailed plan of investigation, acting as a blueprint for testing hypotheses and analyzing data.
• Research design is the sequence of steps taken to ensure relevant data collection for objective analysis of research problems.
• Research design enables researchers to test hypotheses, reaching valid and objective conclusions about the relationship between independent and dependent variables.
• The selection of a research design depends on the purpose of the investigation, types of variables, and conditions of the research.

Purpose of Research Design

• The purpose is to maximize relevant information about the research problem while minimizing costs.
• Research design serves two main functions: answering research questions and acting as a control mechanism.
• Research answers research questions objectively, validly, and economically.
• It aids researchers in data collection for hypothesis testing, variable treatment, manipulation methods, statistical analyses, and potential answers.
• Research design enables researchers to draw valid and objective answers to research problems through related steps.
• Research design acts as a control mechanism, helping researchers control unwanted variances.
• Scientific investigations involve experimental, extraneous, and error variances.

Experimental Variance

• Experimental variance is produced in the dependent variable through manipulation of independent variables by the experimenter.
• Researchers aim to maximize experimental variances for valid and objective data.
• Maximizing variance involves designing experiments where different experimental conditions are as distinct as possible.
• The higher the difference between experimental conditions, the higher the experimental variance.
• Experimenters maximize experimental variance.

Extraneous Variance

• Extraneous variance, or control variance, is produced by extraneous or relevant variables.
• Experimenters control relevant variables, aiming to eliminate variances caused by them.
• Proper control of extraneous variables leads to the elimination of extraneous variance.

Error Variance

• A function of research design is to minimize error variance.
• Error variance includes variances or variabilities in measures due to factors not controllable by the experimenter.
• Factors causing error variance relate to individual differences, errors of measurement, or temporary emotional states.
• Error variance is self-compensating, sometimes positive, sometimes negative.
• Error variance tends to cancel out via repeated measurements, with the mean approaching zero.
• Error variance is unpredictable due to its basis in random errors, differing from systematic variance.
• Minimizing error variances serves two purposes: improving reliability and strengthening external validity.
• Systematic variance is more likely to show significance when error variance is minimised.
• Error variance can be minimized by controlling experimental conditions; uncontrolled conditions increase error variance.
• A good research design seeks results that avoid 'confounding,' where multiple variables change concurrently.
• Confounding introduces vagueness in stating cause-effect relationships.

Criteria of Research Design

• Research design serves as data discipline.
• Criteria exist to distinguish weak research designs from strong ones.
• Criteria include the capability to answer research questions appropriately.
• Good research has the capability to control for the effect of variables.
• Generalizability is key to a useful research design.

Answering Research Questions

• A good research design answers research questions adequately.
• Weak designs do not adequately test the hypothesis.
• Matching irrelevant variables in experimental and control groups results in a weak design.
• Using a two-group randomized design when more groups are needed results in a design that is less able to answer research questions adeqautely.
• Choosing an inappropriate design, such as a two-group randomized design for interaction hypotheses, proves inadequate.

Control of Variables

• A good research design controls the effects of extraneous variables, which have the potential to influence dependent variables like independent ones
• Independent extraneous variables are uncontrolled variables.
• Failure to control extraneous variables leads to a weak design.
• Randomization is considered a top technique.
• Three phases of randomization include random subject selection, random assignment to groups, and random treatment assignment.
• Random assignment of treatments is used when random subject selection is impossible.
• Designs that control extraneous variables are considered the best, increasing the internal validity of the research.

Generalizability

• Generalizability is the external validity of the research, referring to how well results can be applied to other subjects/conditions.
• Designs that allow result generalization are considered good.
• Generalizability depends on technical matters and broader basic/applied research problems.
• Generalizability is secondary in basic or fundamental research.
• Applied research prioritizes generalizability, aiming to apply results to other people.

Basic Principles of Experimental Design

• Experimental design underlies experimental investigation.
• Experimental design is a sequence of steps allowing objective analysis of data that is used to determine cause-effect relations between different groups.
• An independent variable is manipulated by the experimenter directly or through variable selection, so that effects can be monitored.
• A dependent variable changes according to the change in the independent variable.
• Basic principles of experimental design: replication, randomization, and local control.

Replication

• Replication combines duplication and repetition.
• Replication involves deliberate experiment repetition on a different set of participatns.
• Replication validates previous studies, providing an estimate of experimental error.
• Experimental error can be used to measure the significance of observed differences
• Experimental error results from faulty design, measurement, etc.
• Replication should not be confused with multiple measurements.

Randomization

• Randomization is the second basic principle of experimental design, making the test valid.
• Each statistical test relies on the assumption that observations are independent.
• Independence is maintained when subjects or treatments are randomly assigned.
• Randomization ensures independent observations, validates statistical tests, and controls for extraneous variables.
• Randomization controls extraneous variables that would otherwise remain uncontrolled.
• Randomization can be difficult, especially with organismic variables.
• A middle ground between complete randomization is suitable for some cases.

Local Control

• By local control is meant balancing, blocking and grouping.
• Local control is part of the entire design.
• Grouping assigns homogeneous subjects or units into a group for differential experimental treatments.
• Blocking assigns experimental units to blocks by homogeneity.
• Balancing ensures grouping, blocking, and assignment of experimental units is balanced.
• A statistically sound experimental design must possess the property of local control.