The Measurement of Height
Hensley, Wayne E., Adolescence
Height appears to be an important physical attribute. In the social realm, for example, mate selection seems to rest on the male-taller principle (Beigel, 1954; Cameron, Oskamp, & Sparks, 1977; Gillis & Avis, 1980; Wilson, 1989), and being short poses some definite handicaps for males in dating relationships (Hensley, 1994). Among males, height is a factor in occupational opportunities (Hensley & Cooper, 1987), and there is evidence that this extends to later job advancements for both males and females (Egolf & Corder, 1991; Hensley, 1993). Putting the matter into dollars and cents, each inch of height seems to increase one's salary by $600, at least for males (Frieze, Olson, & Good, 1990). Moreover, the financial advantage remains at 4-6% of starting salary even after controlling for job motivation, age, family background, and educational attainment (Loh, 1993). What makes the economic figures so interesting is the consistency of the pattern. As long ago as 1968, Deck reported an 8-10% premium for the starting salaries of academics based solely on height. Moreover, the link between height and salary is not confined to the United States. Steckel (1983) found a strong and highly consistent relationship between height and per capita income in over 30 countries.
The advantages of height begin early. It has been shown that shorter schoolchildren are held back a grade more often than are their taller counterparts (Holmes, Hayford, & Thompson, 1982; Holmes, Thompson, & Hayford, 1984). In addition, self-ratings of perceived competence have been found to be related to height among adolescents (Nottelmann & Welsh, 1986). Thus, height appears to be a physical attribute that has considerable ramifications across a person's life span. Brevity may be the soul of wit, but being short is clearly no laughing matter.
As social scientists probe ever more deeply into the variables that govern the pattern of our lives, it is important that the measurement of those causal agents be precise. To that end, height has received its rightful share of attention. Research has indicated that the correlation between self-reported height and measured height ranges from .93 to .99 (Cameron & Evers, 1990; Charhey, Goodman, McBride, Lyon, & Pratt, 1976; Crockett, Schulenberg, & Petersen, 1987; Lass, Andes, McNair, Cline, & Pecora, 1982). Investigators have stated that self-reports of height provide "valid approximations" (Biro, 1980, p. 46), are "remarkably accurate" (Stewart, 1982, p. 308), that the errors are "very small" (Palta, Prineas, Berman, & Hannah, 1982, p. 229), and the biases are "small and inconsequential" (Palta, Prineas, Berman, & Hannah, 1982, p. 230). It may be that these laudatory comments were based on the fact that self-reported height yielded group means that were close to the sample mean (Tienboon, Wahlqvist, & Rutishauser, 1992). In sharp contrast, the present study focused not on measures of central tendency, but on individual errors.
Usually, glowing assessments and correlations of the magnitude reported above are sufficient to satisfy all but the most finicky critics. Only when it is realized that small variations in height may produce large differences in life outcomes does the measurement issue take on its full importance. For instance, Wilson (1968), using four groups, reported that perceived status increased as perceived height increased. The point to note is that the total difference in perception of height across all four groups was a mere 2.45 inches. Dannenmaier and Thumin (1964) also examined height and perceived status with four different groups that differed by only 4.5 inches. More recently, Hensley (1993) has shown that the academic rank of four groups of university professors was associated with height, which varied by only 1.06 inches. Thus, examination of the literature in this area makes it clear that David-and-Goliath differences are not necessary to affect life course. …