THE JOURNAL OF COMMWNICATION Vol. 19, December 1969, p. 317-324 A STUDY OF AMERICAN NEWSPAPER READABILITY TAHER A. RAZE Abstract This paper is based on a study of American newspaper readability in metropolitan and non-metropolitan areas. The results indicated that there was a significant difference between front page readability level of metropolitan and non-metropolitan newspapers. The metropolitan newspaper was easier to read. There was a significant difference in readability level between metropolitan and nonmetropolitan newspaper articles. With the exception of the local news, the non-metropolitan means were higher for each of these classifications. There was no significant difference between metropolitan and non-metropolitan Associated Press articles. However, there was a significant difference between United Press International metropolitan and non-metropolitan articles. United Press International articles in non-metropolitan papers were more difficult to read. Any reader knows he can understand some materials more easily than others. This level of readability significantly affects the reader's choice of reading material, reading speed, and degree of comprehension. The two primary variables that influence the nature of the reading material are style and subject matter. Style includes such factors as vocabulary level, method of sentence and paragraph construction, and format. Variables attributable to the reader are individual interests, educational experiences, and intelligence level. It would be an arduous task to obtain measures of all the variables affecting the readability of a particular passage. Most methods for measuring readability have concentrated on measuring word difficulty and sentence length. While this approach neglects such factors as the complexity of ideas, sentence and paragraph construction, and individual differences, Taher A. Razik is Associate Professor of Education at the State University of New York at Buffalo. He received his PbD. from Ohio State University in 1W3. During this academic year, 1969-70 he is spending his sabbatical leave as a consultant to the International Institute for Educational Planning, UNESCO, Paris. In this capacity he is developing a system approach to Education for Developing Countries.
318 The Journal of Communication, Vol. 19, December 1969 it has proved to be a relatively valid measure of readability, if the teachers, reading specialists, and librarians judgements are used as the criteria [5]. The concept of readability has generated a great deal of interest among educators. J. B. Conant [2] has suggested that all students who have successfully completed the ninth grade should be able to read and understand the front page of a newspaper. As the newspaper is one of the most important means of mass communication, this recommendation is certainly appropriate. However, Conant s suggestion presumes that newspapers are written, at most, on the ninth grade reading level. The author questioned this presumption and undertook a study of newspaper readability level. In reviewing the literature, no directly relevant studies were found. Specifically, the problems considered were : what is the readability level of the front page of American metropolitan and non-metropolitan newspapers? Does readability level vary for different types of articles and for articles prepared by various news agencies? Hypotheses The hypotheses tested were: 1. There are no articles on the front pages of American newspapers written above the ninth grade reading level. 2. Front page newspaper articles pertaining to space and international and national non-political news are written above the ninth grade readability level. 3. Front page newspaper articles related to crime, tragedy, and features are written below the ninth grade readability level. 4. There is no statistically significant difference (.05 level) in readability between front page articles of Associated Press or United Press International appearing in metropolitan and nonmetropolitan newspapers. Method Sample The population of American newspapers was defined as those American published newspapers that concentrate on general coverage of local, state, national, and international news. Denominational, tabloid, scandal, race-oriented, professional, and highly specialized newspapers were omitted from the target popu-
Razik: Newspaper Readability 319 lation. The population was divided into two strata: those newspapers serving a metropolitan area and those serving a nonmetropolitan area. A metropolitan area was defined as a city with a population of at least 50,000, as determined by the 1960 census. From the 212 metropolitan cities in the United States, 50 cities were selected by simple random sampling. Ayers Dictionary of Newspapers was utilized to construct a list of all the newspapers published in each of these 50 cities. Simple random sampling was employed to select one newspaper from each. For the non-metropolitan strata, Ayers Dictionary was again utilized to create a list of all non-metropolitan areas with newspapers. Fifty areas were selected by simple random sampling procedures. Generally, each of these non-metropolitan areas had only one newspaper which, therefore, became a sample element. In an area with more than one newspaper, simple random sampling was employed to choose one paper from the area. A newspaper was obtained from each of the 50 metropolitan and 50 nonmetropolitan publishers. All the newspapers were dated within a two week period. Articles on the front pages of each of the newspapers were topically classified into the following categories : crime, Cuba, American economy, feature, international, local, national nonpolitical, national political, state political, state non-political, space, tragedy, and weather.l Articles containing lists of names or numbers, society news, brief situational highlights, advertisements, and articles of less than 100 words were omitted from the sample. Utilizing simple random sampling, six articles per newspaper were chosen from the classified front page articles. Each of these 600 articles was subdivided into continuous 100 word sampling units and one 100 word sample was selected by simple random sampling from each article. Procedure The Dale-Chall formula was chosen to predict readability. Studies [ 11 have indicated that this formula provides reasonably valid results. It is based on two variables: proportion of words not included on the Dale Vocabulary list and average sentence length. The Dale Vocabulary list contains 3,000 words found to This study was conducted during the 1962 Cuban crisis.
320 The JournaE of Communicatio.n, Val. 19, December 1969 be familiar to at least 80 per cent of fourth grade children. To utilize this formula, the number of words in a 100 word sample not appearing on the Dale List are counted. The average sentence length is determined by dividing the number of sentences in the sample into 100. Knowing these two values, the sample readability raw score can be read directly from a table [4] and the corresponding readability grade level interval read from the Dale-Chall correction table [3]. Each of the 600 samples was analyzed in the above manner. Raw score means and standard deviations were computed for the article classifications studied in both the metropolitan and nonmetropolitan newspapers. The t test was used to determine significance of differences between metropolitan and non-metropolitan means. Results The results indicated that there was a significant difference (.05 level) between front page readability level of metropolitan and non-metropolitan newspapers. The metropolitan newspaper was easier to read than the non-metropolitan paper. Considering the specific types of articles studied, there was a significant difference in readability level between metropolitan and nonmetropolitan newspaper articles. With the exception of the local news, the non-metropolitan means were higher than the metropolitan means for each of these classifications. There was no significant difference between metropolitan and non-metropolitan Associated Press articles. However, there was a significant difference between United Press International metropolitan and non-metropolitan articles. United Press International articles in non-metropolitan papers were more difficult to read than those in metropolitan papers. Table 1 presents the raw score means and standard deviations. It should be noted that level of readability and numerical score is a linearly increasing function. A conversion of these mean raw scores to corresponding grade level intervals of reading difficulty revealed that the metropolitan newspaper articles concerning crime, Cuba, features, local news, national political news, tragedy, and weather were written at or below the ninth-tenth grade level. Articles pertain-
Razik: Newspaper Readubility 321 Table 1 Metropolitan and Non-Metropolitan Newspaper Readability Scores Metropolitan Non-Metropolitan Type of Article Mean S.D. Mean S.D. Crime Cuba Economy Feature International * Local * National Non-Political * Political-National * -State Space State Non-Political Tragedy Weather * TOTAL * Associated Press United Press International * 7.25.69 7.77.55 8.32.90 7.08 1.06 8.11.83 7.90 1.10 8.19 1.04 7.77.99 No Data 8.25.93 8.37.96 8.96.69 6.32.63 7.84.18 7.94 1.08 7.98 1.01 8.11 8.72 8.82 6.79 8.41 7.29 8.60 8.76 8.20 8.41 8.35 7.11 7.22 8.09 8.37 8.04.95.35.81.80.69 1.18 1.07 2.38.41.94.97.88.31.54.75 1.03 *t test indicated a significant difference (.05 level) between metropolitan and nonmetropohtan newspapers. ing to economy, space, international news, and state and national non-political news were written at the eleventh-twelfth grade level. In the non-metropolitan newspapers, the number of articles written above the ninth-tenth grade level was even greater than in the metropolitan papers. Only articles about weather, tragedy, local news, and features were written at or below the ninthtenth grade level. Inspection of the total readability grade levels indicated that the least difficult articles to read were those concerning crime, features, local news, tragedy, and weather. Table 2 presents the converted grade level readability scores. Discussion As mentioned in the introduction, Conant asserted that students who have successfully completed the ninth grade should be able to read and understand the front page of an American newspaper. This study indicated that the only articles appearing on the front pages of American newspapers concomitant with a
~ 322 The Journal of Communication, Vol. 19, December 1969 Table 2 Metropolitan and Non-Metropolitan Newspaper Readability Grade-Level Intervals TYD~ of Article MetroDolitan Non-MetroDolitan Total Crime 9-10th grade Cuba 9-10 Economy Feature 9-10 International Local 9-10 National Non-Political Political-National 9-10 ll-12th grade 7-8 9-10 -State No data Space State Non-Political Tragedy 7-8 9-10 9-10 Weather 7-8 9-10 7-8 TOTAL 9-10 9-10 9th-10th grade 7-8 9-10 ninth grader's readability level were articles related to crime, features, local news, tragedy, and weather. One can safely assume that these types of articles were not the entire subject matter Conant believed ninth graders should be able to read and understand. Further, the median level of education completed by persons 25 years and over in the United States is 10.6 years [6]. This implies that, for this sample, some important articles (i.e. international and national news) were written above the reading level of at least 50 per cent of the adult population. Articles on crime, tragedy, features, and local news may be interesting; however, they contribute little to an individual's awareness of national and international affairs. Thus, if the newspaper is to be an effective means of mass communication, either the schools must intensify the teaching of reading skills or the newspapers must lower the reading difficulty level of their articles. Other factors which may be relevant are differences in intellectual abilities and the holding power of schools. Tests of hypothesis four indicate rejection of two parts and acceptance of one part. This hypothesis asserted that there was no significant difference in readability level between articles
Razik: Newspaper Readability 323 appearing on the front pages of metropolitan and non-metropolitan newspapers. The results of Table 1 indicated that a significant difference in readability level existed between metropolitan and non-metropolitan papers. The corresponding converted grade levels revealed that the metropolitan papers were written at the ninth-tenth grade level and the non-metropolitan papers were written at the eleventh-twelfth grade level. No significant difference in Associated Press readability level was found between articles appearing in metropolitan and non-metropolitan papers. However, there was a significant difference between United Press International articles in metropolitan and nonmetropolitan newspapers. There are several possible explanations of the above results. Metropolitan newspapers must frequently compete with other newspapers in the area. Therefore, these newspapers might tend to capitalize on the more exciting news events, i.e., crime and tragedy, for front page headlines and news stories. The rationale would be to attract the potential buyer s attention at a newsstand. This hypothesis could be tested by means of circulation indices. Non-metropolitan newspapers frequently have no competition from other newspaper publishers. Therefore, these nonmetropolitan publishers might not realize the same need to attract customers with the lurid and sensational. Further, the non-metropolitan publisher, being a surveyor of the few local, written, mass communication media, may assume more responsibility toward keeping his readers abreast of international and national conditions than might the metropolitan publisher. Another possible explanation of this difference in readability level between metropolitan and non-metropolitan papers might be attributed to the size differences in the editorial staffs. Metropolitan newspapers generally have large editorial staffs and can, therefore, readily have press agency articles rewritten to lower readability levels. Small newspapers with few editors may have to print press releases as they are received from the agencies. This may also explain the difference in readability level of United Press International articles in metropolitan and non-metropolitan newspapers. However, if this is true, then one would also have expected the same difference in Associated Press articles. Thus, these results indicated the need for a follow-up study
324 The Journal of Communication, Vol. 19, December 1969 that would ascertain the type and frequency of articles being written by each of the major press agencies, and survey newspaper publishers to determine the degree of control placed over newspaper readability level. There also appeared to be readability level differences between the rural non-metropolitan and the suburban non-metropolitan newspapers. Further researchers might thus construct three strata based on newspaper circulation. Within each stratum, future divisions based on geographic area might also be made. If the newspaper is to be utilized to its greatest advantage as a means of mass communication it must be reviewed in the educational as well as the journalistic context. Summa y The average front page article readability level was equivalent to the ninth-tenth grade school achievement level. Significant differences in readability level were discovered between metropolitan and non-metropolitan newspapers. International news items were written at a higher readability level than articles describing local events, crimes, and tragedies. REFERENCES 1. Chall, Jeanne S. Readability: An Appraisal of Research and Application. Columbus, Ohio: Ohio State University, 1958. 2. Conant, J. B. Education in the Junior High School Years. Princeton: Educational Testing Service, 1960. 3. Dale, E. and Jeanne S. Chall. A Formula for Predicting Readability: Instructions. Educational Research Bulletin 27:37-54, 1948. 4. Klare, G. R. A Table for Rapid Determination of Dale-Chall Readability Scores. Educational Research Bulletin 31(2) :4347, 1952. 5. Spache, G. D. Good Reading for Poor Readers. Champaign, Illinois: Gerrard Publishing Co., 1958. 6. U.S. Census of Populations 1960. Detailed Characteristics. United SWs Sturnmay. Final Report PC (1)-1C-Washington, D.C.: U.S. Government Printing Office, 1963, Table 76, p. 2.07