Subject selection bias in psychological research.

Abstract

The purpose of this study was to examine the types of bias operating in the selection of •objects for psychological research with humans. A secondary purpose was to examine the effects of these biases on the generality and significance of findings in the most biased research areas. The Journal of Experimental Psychology and the Journal of Abnormal and Social Psychology for recent years were examined. It was found that 73% of the articles in the J-A.S.P. and 85.7% of the articles in the J.E.P. used college students, chiefly male students enrolled in introductory psychology. Use of the general adult population or even samples representing particular sectors of that population is very rare in J.A.S.P., and not found at all in the J.E.P. It was pointed out that the college population differs markedly from the adult non-college population in many characteristics which are associated with psychological factors, e.g. age, social class, learning ability. It is concluded that the male college student has become the "white rat" of human experimentation and that a broadening of the bases for selecting subjects for experiments is in order. RisumiExamen des types de biais a l'oeuvre dans la selection des sujets (humains) sur lesqueb porte le recherche psychologique et analyse des effets produits par ce facteur sur la generality et h signification des resultats obtenus dans les secteurs de recherche les plus biaises. L'examen 4a Journal el experimental psychology et du Journal of abnormal and social psyckoloiy des annees recentes revile que 73% des travaux publics dans le J.A.S.P. et 85.7% des travaux publics dans le JJE-P. portent sur des etudiants de niveau collegial, surtout sur des etudiants masculins inscrits au corns d'initiation a la psychologic Tres rates sont les etudes du J.A.S.P.-il ne s'en trouve absohnnent aucune dans le J.E.P.-qui portent sur une population generate adulte, ou mSme sur des echantuions de secteurs particuliers d'une telle population. Le fait est souligne que la population de niveau collegial differe sensiblement de la population adulte de niveau non-collegial sur plusieurs points lies a des facteurs d'ordre psychologique (v.g. age, rang social, capacite d'apprentissage). L'eiudiant masculin de niveau collegial serait done devenu, pour ainsi dire le "rat blanc" da travail experimental effectue chez des humains et 0 faudrait elargir les entires president a la •election des sujets d'exptrience.

Measuring health beliefs on the Internet: A comparison of paper and Internet administrations of the Multidimensional Health Locus of Control Scale

“…We will not address the issue of the generalizability of results from traditional psychology student samples here, although this issue has been raised by previous authors (e.g., Smart, 1966). We are concerned, however, with the extent to which our data may feasibly generalize to other Internet samples.…”

“…In psychological IMR at least, these approaches (we include those that involve sending requests to e-mail addresses obtained from Internet databases, such as mailing lists) have prevailed (e.g., Birnbaum, 2001;Browndyke, Santa Maria, Pinkston, & Gouvier, 1998;Buchanan, 2000;Buchanan & Smith, 1999;Coomber, 1997;Corley & Scheepers, 2002;Eichstaedt, 2002;Im & Chee, 2004;Kaye & Johnson, 1999;Krantz et al, 1997;Laugwitz, 2001;Riva et al, 2003;Smith & Leigh, 1997;Szabo, Frenkl, & Caputo, 1996). 1 Interestingly (but perhaps not surprisingly), these studies provide evidence that Internet samples accessed using these methods tend to differ in systematic ways from the undergraduate student samples often encountered in psychological research (for evidence that traditional psychological research relies heavily on undergraduate student samples, see Buchanan & Smith, 1999;Smart, 1966). Thus, Internet samples tend to be more diverse in nationality (or geographical location as an approximate indicator of nationality; Birnbaum, 1999;Buchanan & Smith, 1999;Krantz et al, 1997;Senior et al, 1999), of higher educational level (Birnbaum, 1999(Birnbaum, , 2000, more balanced in terms of gender (Bailey, Foote, & Throckmorton, 2000;Buchanan, 2000;Buchanan & Smith, 1999;Riva et al, 2003;Smith & Leigh, 1997), and broader in age range (Buchanan, 2000;Buchanan & Smith, 1999;Eichstaedt, 2002;Krantz et al, 1997;Riva et al, 2003;…”

Diferenças individuais na aquisição de leitura com um sistema lingüístico em miniatura

“…Afi rmações mais conclusivas sobre esses resultados só poderão ser feitas, entretanto, quando estudos adicionais sobre as diferenças na história e motivação desses estudantes forem realizados. A importância desses achados não deve ser obscurecida pela difi culdade de interpretação, entretanto, considerando-se que a utilização de universitários é uma prá-tica comum em experimentos psicológicos em geral (Cozby, 2003;Sears, 1986;Smart, 1966) e sobre aprendizagem de relações arbitrárias em particular. Os resultados apresentados neste trabalho sugerem um maior cuidado na seleção da amostra e na prevenção de viés que a utilização de estudantes de apenas um curso pode produzir nos resultados.…”

Editorial for the Special Issue

“…Thus, much of what we currently consider as psychological knowledge about adult human beings -from human memory processes and personality profiles, to various forms of emotional expression and sexual behaviour -have been based on university student data. In fact, since the mid-1960s, serious concern has been expressed about this situation, and questions have been raised about how well conclusions drawn from university student data can be generalized to human beings at large (e.g., Higbee, Millard, & Folkman, 1982;Higbee & Wells, 1972;Schultz, 1969;Smart, 1966;Wiederman, 1999;Wintre, North, & Sugar, 2001).…”

“…Thus, much of what we currently consider as psychological knowledge about adult human beings -from human memory processes and personality profiles, to various forms of emotional expression and sexual behaviour -have been based on university student data. In fact, since the mid-1960s, serious concern has been expressed about this situation, and questions have been raised about how well conclusions drawn from university student data can be generalized to human beings at large (e.g., Higbee, Millard, & Folkman, 1982;Higbee & Wells, 1972;Schultz, 1969;Smart, 1966;Wiederman, 1999;Wintre, North, & Sugar, 2001).However, as a group and in their own right, university students are very much worthy of being the focus of research investigations, especially in the current tertiary education environment where institutions are busily striving to establish a competitive niche for themselves in the global education community, and at the same time are increasingly being required to operate like real world business entities with stricter and more transparent accountabilities. In such environments, effective teaching and university student performance have reclaimed their rightful place in the centre of institutional concerns.…”