Content Analysis on the Flint Newspapers¹
Have the existence and activity of the foot patrol had any effect upon general public perception of police? To try to answer that question, we turned to analysis of the Flint newspapers where we expected to see the public image of police most clearly.

One thousand two hundred and five (1,205) articles containing some reference to the police were collected from four Flint newspapers between 1979 and 1981. The four were: The Flint Journal, a daily newspaper, The Flint Voice, a volunteer-run newspaper published every two weeks, The East Village Reporter which also publishes every two weeks, and The Flint Voice, a weekly publication that ceased operations in September of 1981. Of the 1,205 articles, 590 are news articles that mentioned in passing that police were investigating some problem, such as a store robbery, some amount of money taken, and a concluding paragraph that the police investigated the incident. One hundred fifty-four (154) other news articles dealt with crime related topics, had no mention of police, or said police did not investigate. Fifty-one articles, features on nonpolice topics, were related to community perception of crime and safety; for instance, articles on the prevalence of shoplifting or the problem of school vandalism.

Of the 353 articles that specifically related to police activities and did not simply state police investigations, 11 percent covered police performance (137 articles, 1 percent covered police community relations (15 articles), and 2 percent covered foot patrol activity (23 articles). Most police articles discussed the Flint Police Department, its budget, its functions, or its politicking.

When the content of articles is considered, several dimensions of police reporting appear. We will look particularly at the overall evaluative tone of the article (negative, mixed, or neutral evaluation, and positive evaluation of the police), how active the police are described as being (passive, routine, advisory, or active), and the specific police activity drawing attention (community problems, technical activities, speaking engagements, etc.).

Sixty-one percent of the total 1,205 articles contained no overt evaluation of police, which concurs with the fact that half the articles are news articles which are supposed to be neutral. One hundred twenty-three articles (11 percent) offer some negative evaluation of police activity, such as one article challenging a police officer's description of a shooting incident, while 111 articles (10 percent) show police in a positive way, such as "to the cheers of nearby residents, police have made more than 25 arrests . . ."; or another article in which a resident is quoted as saying, "I think the police did a good job, even though they were short--handed." One hundred fifty-eight articles (14 percent) offer mixed evaluations of the police-that is some positive and some negative reactions to police or police activity, such as one article where the prosecutor ruled a shooting incident justifiable and proper and, in the same article, the Flint ombudsman said the shooting was unlawful. Clearly, the press balance their evaluations of the police when they make any assessment whatsoever.

The level of police activity was noted in 928 of the articles at four levels of activity: passive (in 2 percent of the articles where police are described as receiving some reports), routine (in half the cases where police are described as following some procedure such as investigation), advisory (in 8 percent of the articles where police are shown speaking at a community meeting or describing how residents can make their homes more secure), or, finally, active (in 40 percent of the cases police are described as apprehending a suspect, raiding an establishment).

What is the content of police activity from newspaper accounts? Seventy-four percent of the descriptions of police activity concern technical police areas like investigation and analysis. Only 3 percent of the articles mention police activity in community relations problems, while 6 percent mention police as public educators (in speaking or teaching engagements), and 4 percent describe the solution of internal department problems. This flat tone of the newspapers is consistent with the 61 percent of the articles that contain no evaluation and the balance given to presentations.

Do police and safety issues loom very large in the Flint newspapers? We asked how much space is devoted to police articles. Forty-five percent of the articles ran 1 to 10 column inches, while 75 percent ran less than 24 inches. There were some very lengthy articles: 7 percent were over 50 column inches long. On the other hand, of the 45 feature articles on safety and community issues, half contained less than 2 column inches mentioning the police. (An article that filled three columns for three inches each would be a 9 column inch article.)

Another indicator of how visible police activities are to the public is where the newspaper places articles on the police. The first page of the first section is usually reserved for issues of great popular interest--the most newsworthy items. The front page of other sections are perhaps secondary in visibility, so a place there indicates news or features of large interest. While there are other criteria for positioning articles we will assume that the least visible articles are those inside the internal sections of the newspaper, not on a front page.

Are police activities a front-page concern? The answer is "yes," for about a quarter of the articles made the front page of section one, while another 12 percent were found on the front page of an internal section of the paper.

Foot patrol was mentioned in only 65 articles (5 percent of all articles), suggesting that foot patrol is not a prominent or visible part of the Flint area police activity. When foot patrol is mentioned, half the mentions have no evaluative component, being neutral; three articles bad negative descriptions of foot patrol, for instance, where a resident would complain that a foot patrol officer had not been seen recently, or that foot patrol was assigned to the wrong hours for the needs of the neighborhood; 23 had positive evaluations of foot patrol such as a commendation of some effort of the foot patrol officer and eight had some mixed evaluation of foot patrol.

Block clubs and neighborhood associations were mentioned in only 36 articles of the 1,205. Of these, 12 articles did not deal with club activity on safety or juvenile issues, while 24 articles on block clubs did reflect some crime or safety concern of the clubs. At least for newspaper articles, it seems clear that safety crime issues are typically seen as a concern of block clubs.

In 927 of the police-related articles, a geographic area of the Flint region was mentioned: 12 percent of these involved the foot patrol area, although the foot patrol itself was not actually cited. That percentage (12 percent) does not seem out of line with the expected representation of any 14 neighborhoods in a metropolitan newspaper. The 14 foot areas do not receive particularly heavy emphasis in the newspapers.

In order to measure the possible overall impact of the Foot Patrol Program upon the public perception of the Flint Police Department, as shown in the newspapers, we turned to analysis of three dependent variables. Of these, two new variables were calculated from the coded data on the newspaper articles. One new variable was called visibility and consisted of multiplication of the number of column inches devoted to the police in the given article by the placement of the article. For placement, an article on the front page of the first section was coded 2; if it was on the front page of another section, it was coded 1; if it was on any inside page, it was coded .5. So, a two inch article on the front page of the first section would have a score of 4 on visibility. If it were on the front page of another section it would have a score of 2, and if it were on an inside page it would have a score of 1.

Why is visibility of possible importance? If the police face a critical incident and handle it poorly, it often makes front page news. The same incident, handled peacefully, may not be as exciting or newsy and may be found on an inner page. The police efforts to improve their image will rarely be as newsworthy as police errors and problems; hence, we expect police visibility should not increase if the Foot Patrol Program is succeeding and improving the Flint Police Department's activity. The foot patrol activity would not increase the visibility of the Flint Police Department since it might make police-community relations less volatile and hense less worthy of front-page notice. We assume that no news is good news on that front.

In order to devise a sensitive measure of visibility, the 1,205 articles were separated into two groups: we dropped cases that were mainly mentions that the police are investigating after a description of some crime, or an article about safety where the police are not mentioned at all, or articles on the police department's budget and functioning. The articles retained were about such matters as foot patrol, the police department in general and its politicking, about community relations, and editorials on crime, juveniles, and city budgets. These were the places where we expected high visibility of the police would indicate a problem for the police department to deal with. To see whether the foot patrol was making any impact upon visibility, a two-way analysis of variance was run in which any mention of the foot patrol in an article was one of the independent variables, and the timing of the article a second independent variable; all articles that appeared in 1979 or 1980 were considered time 1 article and those appearing in 1981 were considered time 2 articles.

The analysis of variance confirmed our expectation that the foot patrol activity did not increase or decrease the visibility of the Flint Police Department in these areas of performance. Even in the second year of newspaper articles, where there was more mention of foot patrol activity, there was no greater visibility of police on the whole. We suggest that this finding means that the positive impacts of foot patrol which may on occasion receive some ballyhoo, nevertheless do not increase the department's visibility. At the same time problems in police-community relations, or police politicking, or editorials on crime and juveniles may be ameliorated somewhat by recognition of the Foot Patrol Program and its impact in community relations. Of course, we cannot prove that the foot patrol made any difference in these cases; but we can note that there has been no increase in the overall prominence in articles about police and delicate areas of their functioning.

A second new variable was created to look at the overall image of the police in the newspapers, no matter where the article on police occurs. Again we look at those 355 articles that were not just descriptions of investigation but rather look at more sensitive areas of police activity. The new variable was called image of police, as indicated by the evaluation of the police (positive, mixed, or good, as already described), multiplied by the number of column inches of the article plus a constant factor 1. Thus a 10 inch article with a good evaluation of the police would count as 10 times 2 for good = 20 + constant 1, or a net score of 21 for image. The same article with a negative evaluation of the police would be transformed thus: (10 inches times 0 for bad image = 0) + 1, with thus a net image score of 1. The range for image was quite extensive from 1 to 80 with most articles being in the range of 0 to 50 on image.

We assume that readers will notice and remember a long positive article about police more than a short, positive article. Since our scale does not run into the negative, a low score on image would indicate a problem for police in its community relations, while a higher score suggests that police are getting a better press in terms of article length and evaluation. To look at the impact of foot patrol over the years upon the image of police, two-way analysis of variance with interaction was run where the two independent variables were the time period (either 1979-80 or 1981) and the second independent variable is whether or not the article had a mention of foot patrol in it. We expected that foot patrol would have a positive effect upon the image of police, particularly as the Foot Patrol Program became better recognized, that is, over time. The analysis of variance showed a strong impact of the year upon image: at p<.001 level of significance, the mean of police was considerably higher in 1981 than in 1979-80. The effect of foot patrol was not as strong, significant only at p<.1, but it was in the direction we expected: that is, articles that had a mention of foot patrol had much higher image of police than did articles with no mention of foot patrol. There was no two-way interaction between the year of the article and mention of foot patrol in the article upon the overall image of the police. This suggests that the foot patrol articles had a weak positive effect over the duration of the grant but the overall image of the Flint police has strengthened during that time, even in articles that do not mention foot patrol activity per se. The esteem of foot patrol could, of course, be a factor in this more positive image, even when foot patrol activity is not mentioned in the articles.

A third area where foot patrol activity might be expected to affect the picture of the police was the variable of level of police activity. We expected that depiction of police as basically passive, or only doing routine activities, is probably seen by readers as a somewhat mundane and negative picture of police. We expected that imagery showing the police as acting strongly or advising communities would be seen as a more positive picture of police activity itself. A successful foot patrol, if reflected in the press, might show more such activity in articles on police. Hence, two-way analysis of variance was run for the effects of year and foot patrol mention upon the mean depiction of police activity level in the 353 articles dealing with more sensitive areas of police concern. As expected, the year did have a strong effect upon depiction of police activity (p<.01), and the mention of foot patrol did have a moderate impact upon the depiction of police activity (p<.07). The 1981 articles about the police showed the police as being more active on the whole than did the 1979-80 articles. Again, articles mentioning the foot patrol see the police as advising and active rather than just pursuing routine activities or being passive. Again, there was no two-way interaction between year and mention of the foot patrol, so that the depiction of police as active rather than passive appears to be occurring even in articles where foot patrol is not mentioned, but increasing throughout the duration of the Foot Patrol Program.
 

¹The extraction of articles from newspapers was done by Hazel Harden and Donna Hale, the coding done by Harden, and the analysis by Philip Marcus and Catherine Smith.