Exploring the Logic of Experimental Design Statistics Project

Exploring the Logic of Experimental Design - Statistics Project Example bjects design is their ability to establish a relationship between a dependent variable and different factors and the fact that the two designs have the same number of participants in each treatment of an experiment (Gravetter and Forzano, 2011). However, while each participant in a within subject design is subjected to all treatments in an experiment, each participant in a matched subjects design is only subjected to one treatment. Correlation in within- subjects design is done across treatments while correlation in matched- subjects design is done across groups of participants. Similarly, subjects are measured in more than one condition in a within- subjects design while measurements on participants in a matched- subjects design occurs only in one condition of an experiment (Gravetter and Forzano, 2011). An experiment’s purpose is to collect data and establish properties of the collected data towards conclusion on a subject. An experimental design accomplishes its purpose by providing approach for conducting the experiment (Jackson, 2011). One of the advantages of experimental design is its high degree of internal validity due to randomization effect that eliminates biasness and causality. It also has a high â€Å"construct validity,† especially when the constructs can be manipulated (Smith and Mackie, 2000, p. 36). The design’s major disadvantage is however in its low â€Å"construct validity† under ethical considerations (Smith and Mackie, 2000, p. 36). Control refers to the manipulation of an experiment’s dependent variable towards an expected observation. Its purpose is to aide multiple group comparison between observations in experimental and controlled set ups (Jackson, 2011). A confound refers to uncontrollable factor in an experiment. Experimental design is an example of a design that can have three confounds. Some of the ways of controlling confounds are â€Å"use of random sampling,† â€Å"use of equivalent control group,† and blinding of