作者: Chris Blais
DOI: 10.3758/BRM.40.4.961
关键词: Machine learning 、 Computer science 、 Statistics 、 Software 、 Selection method 、 Artificial intelligence 、 Caveat emptor 、 Psychology (miscellaneous) 、 Experimental and Cognitive Psychology 、 Arts and Humanities (miscellaneous) 、 General psychology 、 Developmental and Educational Psychology
摘要: The vast majority of psychology labs rely on prepackaged software applications (e.g., E-Prime) for the programming experiments. These programs are often used stimulus selection, and many use a selection method referred to as random without replacement. We demonstrate how replacement deviates from we detail biases that result. also demonstrate, in simple experiment, these biases, if left unchecked, can influence behavior. Recommendations reducing impact performance when is discussed.