How to interpret the results
What does the Index to Market mean?
How is the number of persons calculated?
Are respondents counted more than once in the number of persons?
What methodology does Griffin Media Research use?
What is the Margin of Error for this research?
How to interpret the results
Here is an example of the results produced by the Griffin Reports software:
1. The universe tells you what group of people the results are describing. This can be Adults 18+, Women 25 - 54, or some other set of criteria. The universe description is shown in red on the results table.
2. The profile is what you want to know about the universe you have selected. The profile can be any item in the database, such as broadcast television cumes, household income, or grocery store. The profile you have chosen is the large, bold print at the top of the results table.
3. The composition percentage tells you what percent of the universe has a particular characteristic. For example, if your profile is cable network cumes, and your universe is women 25-54, then you would interpret the results like this: "among women 25-54, 22.8% watched A&E an hour or more in the past week."
4. The index to market gives you a comparison of the percentage of the universe which has particular characteristic compared to the percentage of the entire market having that characteristic. This gives you an indication of the concentration of that characteristic within the universe. For a detailed explanation of the Index to Market, click here.
5. The estimated number of persons shows you approximately how many people in the market that particular item in the results describes. In the above example, the estimated number of women 25-54 who watched A&E an hour or more in the past week is 16,000.
6. The percent of the market and the estimated number of persons that the universe represents is shown at the bottom of the results table. In the above example, women 25-54 represent 29.8% of the respondents in the survey and represent 70,300 people in the market.
Back to Top
What does the Index to Market mean?
The Index to Market is a convenient way to illustrate the relative concentration of the universe to the total market for a particular item in the database. Let's look at an example:
In this set of results, we are looking at the Items Owned for Men 18-34. We see that the index to market for Jet Ski is 169. What this means is that Men 18-34 are 69% more likely than the market as a whole to own a Jet Ski.
Here is how the math works: The composition percentage for Jet Ski for Men 18-34 is 6.4%. The composition percentage for Adults 18+ (the total market) is 3.8%. You can see the composition percentage for the total market by not selecting a universe or clearing the current universe.
The index is calculated by dividing the composition percentage for the selected item in the universe by the corresponding percentage for that item for the total market and multiplying by 100.
6.4 / 3.8 = 1.69 x 100 = 169
Another way to think about the index to market is as a per-capita measurement. Looking at our example again, one can say that on a per-capita basis, men 18-34 own Jet Ski's at a 69% higher rate than the total population.Back to Top
How is the number of persons calculated?
The number of persons is calculated by multiplying the total number of persons for the survey area by the percent of the total market for each item in the database represents. The resulting number is then rounded to the nearest 100 persons.
For example, if the total population for the survey area is 200,000, and the percent of the market that an item represents, such as those who own their homes, is 73.2%, then the estimated number of persons who own their homes in the survey area would be 200,000 x .732 = 146,400.Back to Top
Are respondents counted more than once in the number of persons?
Respondents are counted only once in the composition percent and the number of persons in the results. If the universe criteria allow for more than one option for a resopndent to be included in the universe, any one respondent will be counted only once, even if the respondent meets more than one of the criteria.
For example, if the universe is cable network cumes, and the selected networks are A&E, Discovery, and ESPN, then a respondent is included in the universe if he watches one or more of those networks. The same is true in the reporting of Combos of stations or networks. If you build a combo in the Griffin Reports software, each respondent that watches or listens to at least one of the stations or networks selected is included in the total for the combo. But no matter how many of the selected networks or stations a respondent watches or listens to, that respondent is counted exactly once in determining the composition percent and the number of persons.
This is one of the greatest benefits to using the Griffin Reports software. The ability to produce "unduplicated" cumes allows you to analyze the cumulative net reach within a specific media or across media. For example, you can demonstrate how many (or how few) additional people can be reached by adding competitive media to your own.Back to Top
What methodology does Griffin Media Research use?
Griffin Media Research uses Population Proportional to Size Sampling methodology for research conducted among consumers unless the specifications for the research require a different methodological approach.
Before beginning the data collection process, Griffin Media Research determines the proportions of the population for each standard age & gender category, for each zip code in the survey area, and for race and ethnicity. During the interviewing process, we ensure that the proportions of the completed interviews match the previously determined proprotions for the age, gender, zip code of residence, race, and ethnicity for the population in the survey area.
For example, if men 18-24 make up 6.3% of the total population in the survey area, we ensure that men 18-24 make up 6.3% of the completed interviews.Back to Top
What is the margin of error for this research?
The sampling error is determined by two factors:
1.The sample size for the report
2.The percentage for an item in the report
The maximum sampling error occurs when an answer is 50%. As an answer approaches either 0% or 100%, the sampling error is reduced.
The following table illustrates the sampling error for answers at percentages between 5% and 50% for a various sample sizes:
Most research results are published simply with the maximum sampling error. To determine the maximum sampling error for a given sample size, use the number under the 50% column in the table above.
