The Spanish adaptation of ANEW (Affective Norms for English Words)

JAIME REDONDO 0 ISABEL FRAGA 0 ISABEL PADRN 0 MONTSERRAT COMESAA 0 0 University of Santiago de Compostela , Santiago de Compostela, Spain This article presents the Spanish adaptation of the Affective Norms for English Words (ANEW; Bradley & Lang, 1999). The norms are based on 720 participants' assessments of the translation into Spanish of the 1,034 words included in the ANEW. The evaluations were done in the dimensions of valence, arousal and dominance using the Self-Assessment Manikin (SAM). Apart from these dimensions, five objective (number of letters, number of syllables, grammatical class, frequency and number of orthographic neighbors) and three subjective (familiarity, concreteness and imageability) psycholinguistic indexes are included. The Spanish adaptation of ANEW can be downloaded at www.psychonomic.org. - The Affective Norms for English Words (ANEW; Brad- nication easier. On the other hand, they make it possible ley & Lang, 1999a) provide a set of normative evaluative to conduct comparative studies with different categories ratings for 1034 words in the English language. The ANEW of stimuli and sensorial modalities. Thus, the words in the complements previous work developed by the Center for ANEW have been used as stimuli in such diverse invesEmotion and Attention (CSEA, University of Florida), such tigations as the analysis of neural correlates of emotion as the International Affective Digitized Sounds (IADS; (e.g., Hamann & Mao, 2002; Lewis, Critchley, Smith, Bradley & Lang, 1999b) and the International Affective & Dolan, 2005), the study of activity in facial muscles Picture System (IAPS; Lang, Bradley, & Cuthbert, 1999). (Larsen, Norris, & Cacioppo, 2003), the effects of emo The theoretical position assumed in these works falls tion on memory in youngsters and older adults (e.g., inside the framework of the dimensional perspective of Kensinger, Briedly, Medford, Growdon, & Corkin, 2002; emotions (e.g., Bradley, 1994; Lang, 1995; Lang, Dhillon, Kensinger & Corkin, 2002, 2003), or in investigations & Dong, 1995). This perspective, already defended by on political cognition and attitudes (e.g., Lodge & Taber, Wundt (1896), is based on work by Osgood (Osgood, 2005). Of the three instruments developed by the CSEA Suci, & Tanenbaum, 1957) and conceptualizes emotion (i.e., IADS, IAPS, and ANEW), to date, only the IAPS has as having three basic underlying dimensions along which been adapted to Spanish (Molt et al., 1999). Although the entire range of human emotions can be arranged. The recent standardized databases of the Spanish language two primary dimensions are valence (which ranges from which include objective and subjective indexes exist pleasant to unpleasant) and arousal (which ranges from (LEXESP, Sebastin-Galls, Mart, Carreiras, & Cuetos, calm to excited). The third dimension, not as often men- 2000; B-Pal, Davis & Perea, 2005) none of them include tioned because of its lesser consistency, is dominance or the dimensions of valence, arousal and dominance as subcontrol (ranging from in control to out of control). jective indexes of affective character.1 The wide range of To evaluate a stimulus in these three dimensions, Lang application of the ANEW, along with the aforementioned (1980) developed a nonverbal pictographic measure: the advantages of this database, made it necessary to adapt it Self-Assessment Manikin (SAM; see Figure 1). In the to Spanish, a language that is spoken by more than 400 valence dimension, SAM ranges from pleasant (repre- million people worldwide. sented by a happy figure) to unpleasant (represented by Apart from the ratings of each word in the three afa frowning figure). In the arousal dimension, the SAM fective dimensions, the American standardization of ranges from excited (represented by a figure with the ANEW includes word frequency (Bradley & Lang, overtly open eyes) to calm (represented by a sleepy fig- 1999a). This is the only psycholinguistic index offered ure). Lastly, the dominance dimension ranges from out by the authors. However, it is often necessary to control of control (represented by a huge figure) to in control other psycholinguistic indexes. For example, if one tries (represented by a tiny figure). to use the words in this database as stimuli in a reaction These normalized stimuli databases (i.e., IADS, IAPS, time (RT) task, one needs to control (as well as frequency) and ANEW) are a powerful work tool in investigation, as for variables such as length and neighborhood because, as they make the replication of results and scientific commu- is well known, they affect RTs (see lvarez, Alameda, & Domnguez, 1999, for a brief review). Even the subjective indexes can be of high interest in the investigation. Thus, Altarriba and Bauer (2004) found that emotion words activate different levels of concreteness and imageability in the conceptual representation system of the subject. This way, emotion words are more imageable but less concrete than abstract words and, otherwise, less imageable and less concrete than the concrete words themselves. Thus, values are given of a series of psycholinguistic variables compiled from the databases LEXESP (Sebastin-Galls, Mart, Carreiras, & Cuetos, 2000) and BuscaPalabras (Davis & Perea, 2005). In particular, objective indexes (number of letters, number of syllables, grammatical class,2 frequency and orthographic neighbors) as well as subjective indexes (familiarity, concreteness and imageability) are included. METHOD Participants A total of 720 participants (560 women and 160 men), psychology students ranging in age from 18 to 25 years old (mean 21.5, SD 1.81), contributed to the final data. The ratings of the words were obtained throughout the 20032004 and 20042005 academic years in different faculties of Psychology (or similar degrees) from several Spanish universities.3 Materials and Procedure The words included in the database were obtained from the translation to Spanish of the 1,034 words in the original ANEW (Bradley & Lang, 1999a). This translation was done by a professional philologist specializing in the English language. Then, the list of words was revised by an English philology professor and psycholinguistics researcher under the supervision of the authors. Once the 1,034 words had been translated, three random forms were created. Thus, each form contained the same 1,034 words, but in a different order. Because the evaluation of such a vast number of words would have been tedious for the subject, each form was divided in eight answer sheets. Each answer sheet contained an average of 129 words which were evaluated by the subject in the three emotional dimensions (valence, arousal and dominance). In order to make the assessments, a paper-and-pencil version of the SAM (see Figure 1) was used. It had been adapted from the ScanSAM used by Bradley and Lang (1999) in their normative study. Therefore, just as in this study, the subjects could evaluate each word (...truncated)