American daily newspapers present millions of punctuated sentences to the public, yet--after an extensive search of literature for similar studies--no comprehensive study of punctuation frequency seems to have been made before this study. It is hypothesized (1) that punctuation appears at predictable rates in the main news section of American daily newspapers, (2) that statistically significant differences exist in punctuation frequency between each of five selected newspapers and--by group--between those of large and small cities, and (3) that statistically significant differences exist in punctuation frequency between 1976, 1966, 1956 and 1946. Five newspapers were selected for the study: The New York Times, The Los Angeles Times, The Chicago Tribune, The Atlanta Journal and Constitution, and The Denver Post. The sample dates are February 29, 1976--the first Sunday following the approval of this study--, and the three decades preceeding it, viz., February 27, 1966, February 26, 1956, and March 3, 1946. Samples included the first page of the main news section of each paper. After preliminary studies to determine an appropriate sample size, it was decided to use all words appearing on the first page of each of the twenty newspapers. Twelve marks of punctuation and twenty-three subcategories of use were enumerated as they occurred for each of the 52,184 words which appear on the first page of the twenty newspapers. These include two uses of the apostrophe, colons, eight uses of the comma, dashes, ellipses, terminal and other use of the exclamation point, three uses of the hyphen, parentheses, three uses of the period, terminal and other use of the question mark, two uses of the quotation mark, and three uses of the semicolon. Frequencies were found within all twenty cells which ranged from no occurrences to 111 commas per kiloword. Ninety-five percent confidence intervals were calculated for each use in each cell. These ranged from ± 12.41% to ± 196%. Using the summary data for all punctuation for each of the cells, mean punctuation frequencies and 95% confidence intervals were calculated for (a) each cell, (b) for each newspaper, (c) for large and small cities, (d) for each decade, and (e) for all data. These means and confidence intervals were found for terminal marks and for intrasentence marks as well as all marks. It was found that all punctuation occurred at the rate of 199 marks per kiloword ± 1.9%, that there were 24 words per sentence ± 4.1%, and that intrasentence punctuation occurred at the rate of 156 marks per kiloword ± 2.3%. Analyses of variance were conducted to determine if any of the differences across cells were statistically significant. The number of words per sentence was significantly different by both decade and newspaper at a probability level of 99%. The frequency of all punctuation, and the frequency of intrasentence punctuation did not change by either decade or newspaper.
Library of Congress Subject Headings
Printing; Newspaper layout and typography; Punctuation
Department, Program, or Center
School of Photographic Arts and Sciences (CIAS)
Malcolm, Andrew, "Punctuation Frequency in Selected American Daily Newspapers" (1977). Thesis. Rochester Institute of Technology. Accessed from
RIT – Main Campus