“…Fourteen slides defining a sevenpoint scale of complexity were selected from each set and given to college students to obtain measures of (a) The variable of stimulus complexity has come in for a good deal of attention in recent research and theorizing on the arousal of behavior (Berlyne, 1960;Dember & Earl, 1957;Munsinger & Kessen, 1964;Vitz, 1966a, b;Walker, 1964), Central to much of this work is the concept of an optimal level of stimulation (Fiske & Maddi, 1961;Leuba, 1955) according to which positive affect is produced by stimuli representing a certain value on the scale of complexity, determined by the individual's normal level of stimulation, or more particularly, by some deviation from this adaptation level (e.g., Terwilliger, 1963). This concept of an optimal level of stimulation clearly dictates a research design embodying systematic variation of stimuli along the continuum of complexity, This feature is conspicuously absent in the earlier work of Berlyne investigating the effects of complexity, as in that of others (e.g., Smock & Holt, 1962) directly inspired by Berlyne's, in which stimuli were typically dichotomized into high vs low complexity, or at best into trichotomous categories. More recently, however, a number of researchers (Munsinger & Kessen, 1964;Vitz, 1966a, b) have constructed sets of stimuli varying over an extended range of the complexity dimension, and scaled on an a priori basis, by defining this variable in information-theory terms and constructing randomly generated stimuli of specifiable amounts of information content.…”