Multivariate Analysis of Typeface Damage Frequencies

Abstract

Abstract To explore possible deterministic factors in typeface damage, multivariate analysis was performed on data developed by Dr. David Crown in his doctoral dissertation on typewriter individuality. Crown observed a slow increase in collective typeface damage with the duration of typewriter use, but found no apparent increase in damage correlated with individual letter use frequencies. These findings seemed peculiar and were the initial motivation for the present study. For multivariate analysis, three groups of dependent variables were defined: those based on letter use frequencies; those based on letter form; and those based on basket or keyboard position. Both letter-use frequency and letter area showed statistically significant positive correlation with typeface damage frequencies. Together, these two variables accounted for 42% of the observed variation in damage frequencies. None of the other variables tested was able to explain any statistically significant portion of the remaining variation. These variables included use frequencies associated with difficult finger movements, vertical extension of letters, basket position, and keyboard row. The 58% residual variation must be combined with additional variation introduced by the form and position of typeface damage. With this amount of variation left unexplained, the deterministic factors identified in this study are of little significance in the interpretation of typewriting comparisons.