Measure Validation
ADSAM® Emotional Response Model Validation Summary

Evaluation of Measures
Scientists attempt to develop an accurate understanding of reality by conducting research and utilizing measures that lead to valid interpretations. Given the imperfect nature of measurement, it is important to remember that no one measure is going to be valid for all times and places and that validity is rather a matter of degree. However, the issue of measurement error is very important and conclusions based on invalid measures will themselves be invalid.

Measurement validity exists when a measure (scale) measures what we think it measures and performs the functions that it purports to perform. There are three major criteria utilized to evaluate how well a scale measures what it is intended to measure. First and often considered most important, construct validity is established by demonstrating a scale is related to a variety of other measures as specified in a theory. Second, criterion validity is established when results obtained on one scale can be reliably compared to those obtained with a more direct or already validated measure of the same criterion (phenomenon). Third, content validity establishes that a measure covers the full range of a concept’s meaning.

ADSAM® Validation
Using these standards established by the scientific community, the ADSAM® Emotional Response measurement system passes tests for validity on all three of the major criteria. Background
Measuring emotion requires a conceptualization of emotion and a method for measuring response. Several authors (Aaker, Batra & Holbrook, Holbrook & O’Shaughnessy,Rossiter & Percy) suggest that a limited set of basic emotional responses exists for all stimulus situations. Psychologist Mehrabian and Russell (1974) identified Pleasure,Arousal, and Dominance (PAD) as the core human emotions and all emotional reactions can be described as combinations of these three fundamental dimensions. Mehrabian provides a comprehensive amount of empirical support of the PAD Model of emotions, demonstrating that not only are these three dimensions necessary to describe the full range of emotions, but also, that any more than three only produces diminishing returns.

The PAD Semantic Differential Scale devised by Mehrabian and Russell (1974) is the most widely used and validated instrument for measuring emotional response. This scale is based on sets of emotion-denoting adjective pairs that tap into the intended emotion (i.e., Happy-Sad is used to assess levels of pleasure). The verbal PAD scales have been used in numerous research studies, providing a generous amount of empirical support.

The PAD Model of emotions (Mehrabian & Russell, 1977) gained support in advertising research, especially for strategy formulation and in copy development (Havlena & Holbrook, 1986; Zeitlin & Westwood, 1986). Holbrook and Batra (1988) found that Pleasure, Arousal, and Dominance mediate the effects of ad content on attitude towards the advertisement and attitude towards the brand. Mehrabian (1987, 1992, Mehrabian and Blum, 1995) used the PAD scales in predicting name desirability, predicting product preference, fitting brand names to products, and predicting the effects of different colors.

However, there are a number of inherent difficulties associated with verbal based scales. First, emotional reactions are instantaneous and automatic and respondents spend time evaluating the words used in verbal scales, distorting the original response to the research stimuli. Second, verbal measures are cumbersome, requiring many different ratings for each stimulus presented in an experimental session. This results in a relatively large database that requires rigorous statistical analysis for interpretation. There is a heavy investment of time and effort, limiting the number of stimuli in research settings. Third, the reliance on a verbal rating system makes it difficult to utilize this methodology in non-English speaking cultures (unless there has been translation and validation) and in populations that are not linguistically sophisticated (children).

Functional non-verbal measures are preferable, giving researchers the ability to tap into consumers’ true feelings by bypassing the evaluative processes that distort verbal measures. Hence, Lang (1980; Hodes, Cook, & Lang, 1985) devised the picture-oriented Self-Assessment Manikin (SAM) to directly address these difficulties. SAM was originally implemented as an interactive computer program, and later was expanded to include a paper-and-pencil version for use in groups and mass screenings. SAM ranges from a smiling, happy figure to a frowning, unhappy figure when representing the Pleasure dimension, and ranges from an excited, wide-eyed figure to a relaxed, sleepy figure for the Arousal dimension. The Dominance dimension represents changes in control with changes in the size of SAM. A large figure indicates maximum control in the situation.

Affective reports of the amount of Pleasure, Arousal, and Dominance experienced while processing emotional stimuli can be directly assessed using SAM and plotted on ADSAM® maps with emotion denoting adjectives for analysis and interpretation. In conjunction with the ADSAM® diagnostic procedures, the SAM scales have been utilized in over 600 proprietary studies and 35 academic studies in both the United States and abroad. ADSAM® has been used with different product categories and audience segments. In addition, with the elimination of verbal bias, ADSAM® is easily adapted for use in different cultures.

ADSAM® Construct Validity
First and considered most important is construct validity. Based on the 1957 classic work by Osgood, Suci and Tannenbaum, “Measurement of Meaning”, and later work by Mehrabian, “Framework for a comprehensive description and measurement of emotional states”, ADSAM® is the theoretically related, non-verbal representation of the three PAD dimensions.

Several studies in both the United States and abroad have validated the SAM scales effectiveness in measuring emotional response as proposed by the PAD theory of emotions. In validating SAM’s ability to measure the PAD dimension, subjects were asked to rate verbal stimuli with both the PAD semantic differential scales and the non-verbal SAM scales (Lang, 1980). The correlations between the scales were very strong (Pleasure = +. 937, Arousal = .938, Dominance = +. 660). Near perfect agreement has been obtained in ratings of Pleasure and Arousal for a set of pictures using the semantic differential scale and SAM (Bradley & Lang, 1994).

Morris et al (1992) assessed the effectiveness of SAM in evaluating emotional responses to advertising messages. This was accomplished by using a set of advertisements that had established PAD scores assessed by the verbal semantic differential scales. Scores obtained in the 1988 Holbrook and Batra study are significantly correlated with SAM ratings (Pleasure r = .36, Arousal r = .65, Dominance r = .37).

ADSAM® Criterion (Predictive) Validity
The second standard used to examine a scale, criterion or predictive validity, has been investigated in repeated scientific studies using SAM. Agreement has been obtained in ratings of Pleasure and Arousal for a set of pictures using the semantic differential scale and SAM (Bradley & Lang, 1994). High correlations have also been obtained in tests of emotional response and attitude to the brand and attitude to the ad.

SAM has been compared to previously validated semantic differential scales to measure emotional responses in a variety of situations, including reactions to pictures (Greenwald, Cook. & Lang, 1989; Lang, Greenwald, Bradley, & Hamm, 1993), images (Miller, Levin, Kozak, Cook, McLean, & Lang, 1987), sounds (Bradley, 1994), advertisements (Morris, Bradley, Waine, & Lang, 1995), painful stimuli (McNeil & Brunetti, 1992) and more. In addition, SAM has been used with children (Greenbaum, Turner, Cook, & Melamed, 1990), anxiety patients (Cook, Melamed, Cuthbert, McNeil, & Lang, 1988), analogue phobics (Hamm, Globisch, Cuthbert, & Vaiti, 1991), psychopaths (Patrick, Bradley, & Lang, 1993), and other clinical populations.

ADSAM® Content Validity
Finally, as a non-verbal measure of emotion, ADSAM® does not suffer from the content validity issues of semantic differential scales. In ADSAM® tests, respondents are not judging samples from a pool of “all items”; instead, because of the visual-self-report nature of the device, subjects are judging the complete width and breadth of the dimension.

Morris and McMullen (1993) demonstrated that multiple emotional responses would be found within a single advertisement. In this study, SAM was shown to be useful not only in measuring the full range of emotions to an entire ad, but also to parts of advertisements.

Morris and Karrah (1994) illustrate the stability of SAM in tapping into the PAD emotions elicited by both advertisements and PSAs across product categories. An important finding in this study was the lack of any systematic difference in the responses of males and females, demonstrating no gender differences in using SAM to measure the full spectrum of emotional response generated by advertisements.

Differences obtained in judgments of Dominance suggest that SAM might be more accurate in tracking the subject’s -- rather than the stimulus’ -- feelings of control. SAM allows rapid assessment of what appear to be fundamental dimensions in the organization of human emotional experience, as well as an assessment of the shape of the resulting affective space. Covariation between reports of pleasure and arousal using SAM and responses in physiological and behavioral systems has already demonstrated strong concordance. SAM has been shown to be an effective method for measuring existing feeling states, relating them to other indices of emotional response, and assessing changes due to time, therapeutic intervention, or other processes affecting affective reactions to contextual stimuli.

ADSAM Emotion Adjectives®
Research utilizing subjects ranging from 18-64, has validated the existence of over 153 emotion-denoting adjectives within the three dimensions of PAD emotion space. Subjects are given decks of emotion-denoting adjectives and asked to use SAM to indicate the feeling. Unstable adjectives (i.e. those with little consistency in response) are deleted from the database and stable adjectives are kept. The adjectives are continually tested to assess the stability. The adjectives are regularly used as a frame of reference for analyzing results and guiding strategic decisions.

Cross Cultural Studies
Osgood, May and Miron (1975) collected data from twenty-one communities of diverse languages and cultures and demonstrated that three PAD affective factors are cross-culturally universal. SAM has been empirically used to assess emotional responses in different European countries (Bradley et al. 1993). Results of these studies demonstrate that SAM is reliable and able to detect cross-cultural differences in emotional responses.

Morris et al. (1996) examined the applicability of ADSAM® in a non-English speaking culture to examine SAM’s ability to examine affective responses for global products. The data shows that Hispanic and General market respondents followed similar patterns in their response. Statistical analysis demonstrated there was no significant difference in scores between groups and that SAM did provide comparable measurements to television commercials across cultures.

Morris and Pai (1997) investigated the different emotional responses to global standardized TV commercials with SAM using Taiwanese and Americans 18 to 81 years old. The results showed that there were no significant overall differences in emotional response to ads. These results further support ADSAM® as a method for measuring emotional response globally, across cultures.

Conclusions
The aforementioned research demonstrates that SAM is a viable alternative to well-established verbal measures of emotional response. These studies strongly establish the construct validity (the highest level of validity scientists attempt to achieve in a scale) of the Self-Assessment Manikin. Not only does SAM measure the core human emotions, it also provides a much easier to administer and less cumbersome scale than verbal based measures, allowing many more marketing stimuli to be examined in a single research setting. In conjunction with the Self Assessment Manikin, the ADSAM® diagnostic system provides the necessary tools to gather and interpret the full range of emotional response data to any marketing situation or stimuli.

In addition, the visual nature of the SAM scales and the universality of human emotions enable the ADSAM® system to be utilized to gather and interpret emotional response to global advertisements and cross cultures without additional research or minimal adjustments to the system.

Strictly speaking, one validates the use of the instrument, rather than the instrument itself. Because of this, the developers of ADSAM® work with each client on each questionnaire to insure questions are phrased properly and constructed to uncover the maximum amount of emotional response information.