Methodology
The Canadian Community Newspaper Database - it is the name of both the study and organization that runs it.
Though completely funded by community newspapers, it is an independent company run by a 15-member Tri-Partite Board of Directors, each of whom has an equal vote. There are five representatives from each of the publisher, advertiser and agency categories.
ComBase 2003 and ComBase 2005 were conducted by Maritz Research (formerly Thompson Lightstone).
ComBase 2003/2005 Methodology
With a 2-year sample of over 50,000 interviews in 400 markets and sub-markets, more than 1,000 publications and radio stations, the study is one of the largest in North America.
ComBase employs the ‘Recent Reading’ technique currently in use by readership studies in Canada and around the world. ComBase’s specifications meet and exceed those set other media studies such as NADbank and PMB and, in fact, ComBase was appraised and endorsed by the Canadian Advertising Research Foundation (CARF).
The interview begins with a 3-month screen-in question for a long list of market publications that includes dailies, entertainment, agricultural, automotive, employment, alternative, ethnic and other print publications. All publications identified as read in the past three months are put through the standard recency and frequency suite of questions.
Weekend publications or weekend-like publications (i.e. community newspaper editions whose distributions, formats or editorial are substantially different from the weekday product, but published on a day other than Saturday or Sunday) are measured separately.
ComBase 2003/2005 Interviewing Method
ComBase conducts a 10-minute telephone interview with adults 18+. Respondents are chosen at random and each household, whether it has an unlisted or a newly listed phone number, has an equal chance of being selected. To further ensure the randomness, and therefore the reliability, of the survey, there is no substitution of the respondent allowed.
Interviews are conducted in English, although those with other mother tongues were also interviewed. Bilingual newspapers were also measured, but again, the interviews were conducted only in English.
ComBase 2003/2005 Timing
Interviewing for the 2003 study was conducted between October 2002 and October 2003 and during the Spring of 2004. Interviewing for the 2005 study was conducted between October 2004 and June 2005. There were breaks for Christmas and other significant holidays that might interfere with our ability to find respondents. Market starts were staggered over the interviewing period. The ComBase 2005 Two Year Study is a combination of interviews from each wave.
The sample is released over three waves (one per month), which averages readership measurement over time. One fifth of each wave is released Tuesday through Saturday to maintain balanced day of week readership.
ComBase 2003/2005 Response Rates
ComBase contracts for and achieves a minimum of 40% response rate in each market and sub-market measured. Response rates are the cornerstones of reliable research. We accomplish this difficult task in a variety of ways:
- The interviews are spread out over a minimum of 12 weeks to catch potential respondents at different days and times of the day
- A minimum of 12 repeat contacts must be made to each telephone number
- The newspaper list is filtered through the respondent’s postal code to avoid asking about publications to which he/she has no access
ComBase 2003/2005 Market Definition
In CA’s and CMA’s typically the entire StatsCan geography is measured. Determining the core markets for community newspapers in smaller geographies is difficult because of a variety of factors:
- Whether the publication is paid or free. Though the majority of circulation in Canada is free, the majority of community newspapers are paid products, most of whom distribute to areas comprising hundreds, if not thousands of square miles.
- Whether the publication distributes in a rural or urban market
- How it is distributed (i.e. via Canada Post, carriers, or others)
Publisher distribution areas were gathered at the six-digit postal code level, then, based on minimum penetration levels, grouped into Dissemination Areas (DAs) from which towns, villages and cities were built. All geography used is defined by Statistics Canada and comparable to that used by other studies. Whole DAs were used for all markets.
The market definition inclusion rules and tests are as follows:
Free (39% of members)
- Masthead name is starting point
- Newspaper must achieve 50% household penetration of each DA included, AND
- The circulation in the defined market must represent at least 70% of the newspaper’s total circulation
Paid (61% of Members)
- Start with masthead town name
- Newspaper much achieve 50% household penetration of each DA included
- If not, penetration can be reduced to no lower than 30% to achieve circulation minimum, AND
- The circulation has to equal 50% of the newspaper’s total circulation
For data analysis purposes market information is available at different levels in ComBase:
- CMAs (Census Metropolitan Area) – Population 100,000 or more in an urbanized core
- CAs (Census Agglomeration) – Population 10,000 or more in an urbanized core
- Suburbs (often CSDs)
- Newspaper Market Areas (as determined by distribution and rules above)
- CSD (Census Subdivisions – towns, villages, etc)
- Provinces
What ComBase Measures
Newspaper readership, other media usage and demographic information was gathered. The following details the content:
- Readership of community newspapers, local and non-resident
- Readership of daily newspapers, including nationals
- Readership of other print press including shoppers, agricultural press, alternative publications, etc.
- Amount of paper read
- Frequency of reading
- Preferred community newspaper if more than one read
- Rating of all newspapers read
- Radio listening, TV viewing
- Demographic information about the respondent and household
Sample Size
Since ComBase 2003/2005 was designed to be a two-year rolling sample, the ComBase Board specified a minimum of 75 completions for each newspaper market area. In the ComBase 2005 One Year study the minimum was increased to 100 interviews per market area. This allows for basic newspaper ratings (read last issue) and many fifty-fifty splits such as Household Income over $50M vs. HHI under $50M. For many markets, the study is not designed for combinations of demographics, although now that the second wave of interviews has been added in the 2005 release, many more options will be available for planners. Sample sizes in the two year study are in the range of 300+ per market area.
Margin of Error
The margin of error is a measurement of the accuracy of the results of any survey.
Example: A margin of error of plus or minus 5% means that there is a 95% chance that the responses of the target population as a whole would fall somewhere between 5% more or 5% less than the responses of the sample (a 10% spread).
An important factor in determining the margin of error is the size of the sample. Larger samples are more likely to yield results close to the target population quantity and thus have smaller margins of error than more modest-sized samples.
On an average sample of 150 the margin of error is likely to be in the range of +/- 7.9%. The larger the size of the sample, the lower the margin of error will be. On a sample of 200 the margin of error is +/- 6.9% and on a sample of 300 the margin of error is likely to be in the range of +/- 5.5%.
Contact Info
For questions please contact Kelly Levson, General Manager at kellylevson@combase.ca or 1-800-481-6580.