How demographic variables affect newspaper delivery

How demographic variables affect newspaper delivery

Seamon, Marc

Reliability of newspaper delivery is a serious concern for circulation managers. This quantitative examination of reliability and satisfaction variables for lateness, damage and non-delivery indicates that current industry trends to replace juvenile carriers with “more reliable” adults may be misguided.

Reliability of newspaper delivery is a serious concern for circulation managers because, unfortunately, failed newspaper delivery or other unsatisfactory performance from carriers is not uncommon. The literature has focused on news consumers’ dependency on and uses for the media, and to some extent, their reactions to failed delivery, but an empirical assessment of reliability and satisfaction coupled with a comparison of some factors that may affect those measures is needed.

Commentary and rhetoric from newspaper industry officials and jour nalism scholars raise more questions than they answer. Some say that youth carriers are unreliable, but others defend the little merchant system as more efficient, if not more reliable, than an all-adult carrier force. Some insist that costs can be cut and reliability boosted by having a smaller group of adult carriers deliver longer routes, but are the differences the result of carrier age or route length?

Others say that reliability problems lie with a high rate of carrier turnover. They suggest that the longer a carrier holds a route, the better the service will be because reliability will improve with experience. Are young carriers less reliable than adults? What about carrier experience, he number of customers on a route, route length, or route density? This paper begins to examine how these factors relate to carrier reliability.

Literature review

In the 1940s, Bernard Berelson took one of the first looks at failed newspaper delivery. New York City’s newspaper carriers had been on strike for 17 days, and the city was virtually without paper service when Berelson conducted his study to determine what people miss most about their paper when they can’t get it.1 His results showed such strong reaction among a sample of 60 customers that a new genre of media research was born: dependency theory. As the name suggests, dependency theory indicates a reliance and a need for the media, as compared with a want or desire associated with uses and gratifications theory.2

About 13 years after Berelson’s study, the New York City Newspaper and Mail Deliverers’ Union struck again, leaving the city without papers for 19 days. Penn Kimball, a professor of journalism at Columbia University, used the situation to conduct a study following in Berelson’s footsteps. He surveyed 164 New Yorkers who regularly read a paper, again asking the question: “How much do people miss the newspaper in times of failed delivery?” Kimball found the same strong dependency on newspaper.

In 1998, Clyde Bentley, a doctoral student at the University of Oregon, had a local newspaper provide him with a daily list of customers who had been missed on the delivery route. He then called these people and asked 10 qualitative survey questions. Bentley’s results suggest that people miss their paper very much when they don’t get it. His and other contemporary studies have found the newspaper to be such a powerful component of media dependency theory that terms such as habit and ritual have been used to describe people’s reliance.3

However, Berelson, Kimball and Bentley stopped short of exploring a solution. Because news consumers depend on reliable newspaper delivery, there is good reason to explore ways to provide that service.

Most circulation concerns being discussed today seem to stem from a contentious debate concerning youth versus adult carriers. Several newspapers, including the Portland Press Herald, Maine Sunday Telegram, San Francisco Chronicle, San Francisco Examiner, Belleville Democrat, Toronto Globe and Toronto Mail have eliminated youth carriers or are considering the change.4 Many cite safety concerns for the youth, but most also offer business reasons for the decision.

Some contend that young carriers are not as reliable as adults.5 In 1984, the Toronto Globe and Mail had 3,400 youth carriers who were generating about 4,200 to 4,500 complaints each week. After switching to a streamlined delivery force of adult carriers with longer routes, the papers’ 550 to 580 adult carriers produced 932 complaints weekly.6 Although there were fewer complaints after the switch to adults, the ratio of complaints to carriers increased. It’s impossible to know whether the change was attributable to the adults or to the longer routes, but the memory and learning literature would suggest that as the amount of time spent on task without rest increases, so will the frequency of mistakes because of fatigue.7 It remains to be seen whether this principle extends to newspaper route length and delivery reliability.

When the Belleville Democrat replaced its junior carriers with an all-adult delivery force, the first reason for the change was service, said Frank Duke, the paper’s marketing director.8 Similarly, Bentley listed “oversights by juvenile carriers” as a factor in the daily misses he examined in his research. And in his results, he noted that irate customers often “qualified their statements of irritation with comments of understanding for their young newspaper carriers”.9 No mention was made of adult reliability.

Other practitioners say youth carriers may be more reliable than adults, but add that youth are less efficient than adults. “The down side of hiring adults,” said Jack Butcher, circulation director of the Tampa Tribune, “is that while they give more consistent service, it usually is not as good as that provided by youth”.10

Jack Hertter, circulation manager of the Bloomington Pantagraph, said replacing youth with adult carriers means a less-tailored delivery style. “The customer gets a little less service and pays the same price for it.”11

The literature shows disagreement and uncertainty about the reliability of carriers. Is age a factor? Some in the industry suggest that any advantage in adults may be the result of experience.

“Older carriers can also be rewarded with longer routes, as those with experience can probably finish their deliveries more quickly,” said Jim Jones, circulation manager of the Kenosha News.12 Jones’s assertion would be supported generally by the psychology and learning literature, which suggests that if the factors of massed and distributed practice are equal, task proficiency is directly related to the amount of time spent practicing.13 However, it remains to be seen whether that relationship extends to newspaper delivery reliability and amount of carrier experience.

Data from the Circulation Research Group indicate that young carriers have fewer customers on their routes than adults. Nearly 90 percent of the youth routes examined at 25 northeastern newspapers were composed of fewer than 50 customers. Young carriers also quit their routes more frequently than adults. On an average day, 6 percent of all youth routes were delivered on an emergency basis after the carrier quit without warning, and only about one-third of juvenile carriers kept their routes for one year.14

These factors of age, experience and turnover are generally believed to have an impact on reliability and customer satisfaction, but there is little empirical evidence to support the connection. As the news media increasingly focus on the marketing aspect of audience concerns, circulation managers must ask what they can do to improve reliability. Kenneth R. Todd, circulation director of the Indianapolis Star and Indianapolis News, called on papers to “invest more in circulation.”15 The first step toward that end is to examine the relationships among carrier age, experience, route length, number of customers and reliability of delivery.


H1 Carrier age is negatively correlated with incidents of unreliable delivery. This hypothesis tests the notion that young carriers are not as reliable as adults.

H2 Amount of experience as a carrier is negatively correlated with incidents of unreliable delivery. This hypothesis tests the notion that paper carriers achieve better reliability as they gain experience on the route.

H3 Time spent delivering daily routes is positively correlated with incidents of unreliable delivery. This hypothesis tests the time aspect of the notion that as fatigue mounts, reliability decreases.

H4 Route size, in number of customers, is positively correlated with incidents of unreliable delivery. This hypothesis tests the size aspect of the notion that as fatigue mounts, reliability decreases.

H5 Route length is positively correlated with incidents of unreliable delivery. This hypothesis tests the distance aspect of the notion that as fatigue mounts, reliability decreases.

H6 Urban route density is positively correlated with incidents of unreliable delivery. This hypothesis is based on excuses given by negligent carriers who suggest that errors are more frequent in highly populated neighborhoods where subscribers are closely mingled with non-subscribers.

H7 Customers with a paper box report better service than those without one. This hypothesis tests the notion that issuing paper tubes (boxes) is a good way to increase service, both through better reliability and decreased damage.

H8 There is a significant difference between delivery reliability of morning and evening routes.


This project involved a cross-correlation analysis to identify possible relationships among newspaper circulation reliability variables and delivery route demographic variables. It measured differences between circulation reliability variables compared on the basis of morning versus evening delivery and tube versus porch delivery.

Two daily newspapers, one morning and one evening, each with circulations of about 20,000, were selected for this initial study. Both papers were printed and delivered by the same company in the same city. Hopefully, selecting test papers in this way eliminated some of the variance that may have been encountered had the samples been taken from papers published in separate areas.

Route demographic variables

A route master report was obtained from the circulation department of each paper. This document is a complete list of all paper routes in the circulation area. It indicates which carrier currently is delivering each route, how many customers are on each route, and the location of each route within the circulation area.

The number of routes shown on the route master report was misleading because a single carrier often has two, three or even more routes. The carrier, of course, treats the multiple routes as one. Another problem occurred when carriers had other people helping them with routes. In some instances, a carrier would take half the papers and go one way to deliver while a friend or relative took the rest and delivered in another part of town.

Pending telephone confirmation, multiple routes held by the same carrier were lumped together to represent what they really were: one, big route. Likewise, if a carrier delivered only half the papers indicated on the route master report, the appropriate correction was made.

When both of these problems were corrected, and an accurate picture of the route sizes and carriers was established, a stratified sample of carriers was selected based on location and the number of customers comprising routes.

After the routes necessary to construct a stratified sample were identified, the carriers holding those routes were contacted by telephone. Each carrier first was asked to confirm the number of customers along the route. The number of customers the carriers reported sometimes was different from the number indicated on the route master report because they continually acquire or lose subscribers. It takes time for these changes to be reflected in the route master report.

Five information variables were collected from each carrier during the telephone interview. The carriers were asked: 1) how many customers comprised their weekday routes, 2) how long it took them to deliver their routes on an average day, 3) the length of the route in kilometers, 4) how many months they had held the routes, and 5) their ages in years. A sixth variable was calculated by dividing the number of customers on each route by its length in kilometers. This variable, labeled density, was meant to quantify the urban or rural quality of each route and measure changes in relation to that characteristic. Some country routes were long and spread out, while city routes often were short and dense.

Reliability variables

After these six variables were collected from the carriers, another report, called a customer list, was obtained from the circulation department. The customer list contains the name, address and phone number of every customer on each route. It also indicates each customer’s subscription type (seven-day, six-day, five-day, weekend-only or Sunday-only). Weekend and Sunday-only customers were eliminated from the lists because it was not valid to compare them with customers who received the paper all week. Only customer lists for the routes selected in the stratified sample of routes were obtained.

Households were selected randomly from the customer list of each sampled route and contacted for a telephone survey. Each route was considered separately. That is, a phone survey was conducted for each route, and the results were paired with that route’s demographic variables (the information collected from the carriers).

The number of households surveyed along each route varied to match the route size. For routes consisting of fewer than 100 subscribers, survey responses were collected from 10 people. This method allowed the percent sampled to increase for smaller routes, which demand a larger sampling percentage to ensure a valid measure. Using this method, a full 50 percent of the households would be surveyed from a route containing 20 customers, while a smaller proportion – about 17 percent – would be sampled from a route of 60 subscribers. For routes composed of more than 100 subscribers, a sampling quota of 10 percent was collected. Using this method, a large route of 320 customers would be represented by 32 respondents. This seemed to be the most reasonable method to collect samples for comparison of populations varying in size from 11 to more than 400.

These questions were asked during the phone interviews with customers:

1. How many times in the last two months has your newspaper been delivered more than three hours late?

2. How many times in the last two months has your newspaper been delivered to you damaged, like being wet or torn?

3. How many times in the last two months has your newspaper not been delivered at all on a given day?

For each of the above questions, those respondents reporting incidents of lateness, damage or non-delivery also were asked to quantify their satisfaction with the respective performances using a 10-point Likert scale.

Lateness, damage and non-delivery were assessed for the last two months because that was the shortest period any of the sampled carriers had been with the newspaper. To ask respondents to report performance for a longer period, say the last six months, would have compromised validity if their carrier had been on the route for only three months.

The data were processed by calculating item means for the incidents of lateness, damage and non-delivery as well as for the respective Likert-scale satisfaction responses. These means were calculated on a route-by-route basis, with each route producing item means for the six reliability questions.

Comparing the variables

The six demographic variables collected from the carriers were correlated by route using Pearson’s r with the item means of the six circulation variables collected from the customers. This was done to search for patterns among variables such as route size, density and carrier age and incidents of lateness, damage and non-delivery of the newspaper.

Finally, subscribers were asked if they had a paper box. This yes/no question was posed to determine whether people with paper boxes receive more reliable delivery than those who don’t have a box. Incidents of lateness, damage and non-delivery reported by box customers were separated from the responses of those without boxes.


A total of 722 subscribers was interviewed along 54 newspaper routes. On average, subscribers reported .625 incidents of late delivery (more than three hours late) in the last two months. Respondents also reported .725 incidents of damage and .359 incidents of non-delivery (no paper at all). Using a zero-to– 10 Likert scale, mean satisfaction ratings for the degree of lateness, damage and non-delivery were 9.25, 9.33 and 9.23 respectively.

Hypothesis 1 was not supported. Age was not significantly correlated with reliability. As predicted, the correlation coefficients for lateness, damage and non-delivery were negatively associated with carrier age. But at -.245, -.215 and -.202, respectively, they were not significant at the 95-percent confidence interval (p= .07, .12 and .14).

Hypothesis 2 was not supported. Experience as a carrier was not significantly correlated with reliability. Again, the associations were in the direction predicted by the hypothesis, but they were too weak to consider real.

As is illustrated in Hypothesis 1, age was not correlated with reliability. However, age was strongly correlated with several other measures, including number of customers, route length in minutes and kilometers, and route density. (see Table 1) What that means is that older carriers had larger routes that contained more customers, took more time to deliver, were spread out over longer distances and were more rural. Therefore, comparing route length, number of customers and route density with reliability was, in effect, comparing age with reliability. To control the variance introduced by the age factor, the data were divided into adult (18 and older) and juvenile (17 and younger) age groups. The correlations for hypotheses 3, 4, 5 and 6 then were calculated using both the original and the age-sorted data sets. There were 23 youth routes and 31 adult routes.

Hypothesis 3 was supported in the youth routes but unsupported in the adult routes. For juvenile carriers, route length in minutes was positively correlated with incidents of non-delivery (r= .529 p= .01). No similar effects were observed in adult routes or in the unsorted data.

Hypothesis 4 was not supported. Route size in number of customers was not related with reliability.

Hypothesis 5 was supported in the youth routes but unsupported in the adult routes. For juvenile carriers, route length in kilometers was positively correlated with incidents of non-delivery (r= .443 p= .04). No similar effects were observed in adult routes or in the unsorted data.

Hypothesis 6 was not supported. Route density (number of customers divided by length in kilometers) was not related with reliability.

Hypothesis 7 was supported. Customers with a paper box reported significantly fewer incidents of lateness, damage and non-delivery than did customers without a box.

Hypothesis 8 was supported. Morning customers reported significantly fewer incidents of lateness and damage than evening. Morning and evening deliveries were not significantly different when compared by incidents of non-delivery.


For circulation managers seeking to improve delivery reliability, this research suggests that the current trend of eliminating juvenile carriers in favor of adults probably is a mistake. Age was not significantly related to reliability. There were some amazingly fastidious carriers of all ages, as well as some downright lousy ones.

However, the examination of age did reveal that some good can be accomplished by keeping youth routes short. To maximize reliability, circulation managers should not assign long-distance or very time-consuming routes to their younger carriers. Among carriers 17 and younger, incidents of nondelivery increased with time and distance of the routes. Adult carriers showed no such susceptibility to time or distance, so circulation managers should assign them the longest routes.

However, a route with a lot of customers is not necessarily a poor choice for a young carrier. Even among the youth-route data set, number of customers was not related with reliability. Remember that route length in minutes and kilometers was, however, related with reliability.

It should be noted that in the youth routes, the negative correlation between reliability and the route length factors of time and distance does not represent support for the notion that youth are less reliable than adults – that assertion was tested and unsupported in Hypothesis 1. In none of the data sets (youth-only, adult-only or unsorted) was age significantly correlated with reliability. It was not age, but rather route length that was compared with reliability in Hypotheses 3 and 5. The length factors were significantly correlated with reliability in the youth routes, but not in the adult or unsorted data sets. And as Hypothesis 4 suggests, reliability was not significantly correlated with the number of customers in any of the data sets. Those findings suggests ways to improve performance in youth routes, but they do not indicate that youth are less reliable that adults.

A factor that did not seem to affect performance for carriers of any age was the density of their routes. No matter whether the routes were highly compressed in the city or spread out in the country, reliability and density were unrelated.

However, as practical wisdom might suggest, issuing a newspaper box was an effective way to improve reliability. The 290 customers with boxes reported much lower rates of damage and non-delivery than did the 434 customers without them. Those results dovetail with what one would expect a paper box to do. Boxes appear to protect against weather damage and seem to give carriers a visual cue that reduces non-delivery “misses.” The advantages in decreased damage and non-delivery were so significant that circulation managers should be encouraged to issue a box to every customer who will accept one. Of course, if a carrier is simply late, a paper box won’t help, and accordingly, there was no significant difference between the groups based on lateness.

Paper boxes or tubes are issued to any customer who requests one. Sometimes they are mandated for customers who live at the end of long country lanes or private drives. In these cases, the tubes are placed along a more-accessible main road just as mailboxes sometimes are.

The significant differences between morning and evening delivery indicated in Hypothesis 8 probably are linked to several causes. Likely, the most powerful among these is the fact that there is more time to fix press malfunctions that occur overnight than there is to repair those that occur in the afternoon. Also, many carriers say that early morning traffic is lighter and less distracting than afternoon rush hour, during which many evening papers are delivered. These explanations for the differences would be consistent with the findings that lateness and damage were different but that non-delivery was not. Also, the ratio of adults to youth was higher for the morning routes than the evening routes, so some of the variance likely is the result of age as well.

Although no data were collected to measure their impact, polyethylene bags also seem to be remarkably effective in reducing damage, at least in the eves of respondents offering qualitative remarks about reliability.

Finally, circulation managers should know that this research found no relationship between experience as a carrier and reliability. The results found here regarding Hypothesis 2 suggest there is no reason to expect that a newspaper carrier who has held a route for several years will be any more reliable than one who started last month, regardless of age. More than any other finding, that result perhaps underscores this research. Reliable delivery seems to be more strongly related to the work ethic of the carrier than many demographic qualities of either the carriers or their routes. Perhaps good carriers are earmarked by traits such as responsibility. To be sure, they come in all ages and deliver all types of newspaper routes.


1. Bernard Berelson, What Missing the Newspaper’ Means. Communication Research, P.F. Lazarfeld and F.N. Stanton, eds.. New York: Harper and Brothers, 19481949, pp. 111-128.

2. August E. Grant, Youda Zhu, Debra Van Tuyll, Jennifer Teeter, Juan Carlos Molleda, Yousef Mohammad and Lee Bollinger, Dependencyand Control. Paper presented at the Association for Education in Journalism and Mass Communication annual convention, Baltimore, Maryland, 1998.

3. Clyde Bentley, 50 Years Later: `What it Means to Miss the Paper,’ Berelson, Dependency Theory and Failed Newspaper Delivery. Paper presented at the AEJMC annual convention, Baltimore, Maryland, 1998.

4. Don Sneed, The Case for Youth Carriers. Editor & Publisher, July 19, 1986, pp. 44, 35; Debra Gersh, The Move to Adult Carriers. Editor & Publisher, November 8, 1986, pp. 26-29, 39; Joe Strupp, San Francisco Youth Carriers Fight to Keep Jobs. Editor & Publisher, September 10, 1994, pp. 28-29; Sara Rimer, Youthful Newspaper Tradition is Ending. New York Times, December 15, 1996, Sec. A, pp. 30.

5. Bentley. op cit.

6. Sneed, op cit

7. Ulric Neisser, Memory Observed, Remembering in Natural Contents. San Francisco: W.H. Freeman and Company,1982; R.A. Magill, Motor Learning Concepts and Applications. Dubuque, Iowa: Wm. C. Brown Publishers, 1989.

8. Debra Gersh, The Move to Adult Carriers. Editor & Publisher, November 8, 1986, pp. 26-29, 39.

9. Bentley, op cit.

10. M.L. Stein, Youth vs. Adult. Editor & Publisher. July 7, 1990, pp. 22-23. 11. Stein, op cit

12. Gersh, op cit.

13. N.L. Perkins, The Value of Distributed Repetitions in Rote Learning. British Journal of Psychology, 1914, pp. 253-261; E.C. Cooper and A. J. Pantle, The Total Time Hypothesis in Verbal Learning. Psychological Bulletin, 1967, pp. 221-234.

14. C.H. Favor, Collapse of Home Delivery is Threatened. Editor & Publisher, December 5, 1987, pp. 68, 45.

15. Anonymous, Reducing Carrier Turnover. Editor & Publisher, May 28, 1988, pp. 24, 39.

by Marc Seamon

The author is chief of the Marshall County Bureau of the Wheeling (West Virginia) News Register. He has a master’s degree in journalism from the Reed School of Journalism at West Virginia University in Morgantown.

Copyright E.W. Scripps School of Journalism Winter 2000

Provided by ProQuest Information and Learning Company. All rights Reserved