6-week Cume/Pref reports menu
You can review Cume and Preference information in the form of a 6-week trend by either Discrete Demo (percentage or raw counts), Rolling Average, Combined Demos (raw counts), or Combine Both Boxes.
Discrete Demo prints a separate page for each demo cell you're testing (Men and Women are combined, although you can filter for either). Here's an example of the 6-week Trend 18-24 Female page:
6-week Discrete Demo Cume/Pref report / percentages
This same report can also be run in Raw Counts mode. Instead of calculating the percentage of each week's sample as a percentage, the actual respondent counts are displayed:
6-week Discrete Demo Cume/Pref report / raw counts
This cume/pref report lets you specify a date, demo and gender range, and also how many weeks of results you wish to roll together. The results are still displayed in 6-week cycles, but each cycle may be comprised of anywhere from 1 to 9 weeks worth of data.
Cume/Pref report filter
For example, in the screen shot above, we are rolling together 4-weeks of data, in each cycle of the report. Here's the report that we have generated:
Rolling Cume/Pref report
Each weekly column in the report is actually comprised of that week, as well as data from the 3 prior weeks. For example, the column of data for 1/13, also includes the data from 1/6, 12/30 and 12/23. The data in the column 1/6 also includes data from the weeks of 12/30, 12/23 and 12/16. By rolling 4-week's worth of data together into each cycle of this report, we have essentially quadrupled the sample size for each cycle, and will thus minimize the inherent wobble in each week's sample. This is similar to the rolling trend methodology utilized by Arbitron; as a new month of sample is added to a 3-month trend, the oldest month of sample falls off.
Raw Counts Combined
This report allows you to combine demo/gender cells, and create one report including these cells. In the example below, we have created a 6-week trend report for 18-54 Men and Women. (18-54 Persons)
6-week Cume/Pref Raw Counts report
If you want to generate a cume/pref report that includes two or more non-contiguous demo cells, you will need to run the Raw Counts/Combined report (otherwise the Discrete Demos reports will automatically print one demo cell per page for you).
Combine Both Boxes
This report combines the mentions by respondents of stations in both boxes of your screener (Cume and Preference). This creates an aggregate count for each station, similar to total gross impressions. This report is particularly useful if you are using the two boxes on the screener to collect First Favorite and Second Favorite information. In that case, you would merely want to combine the mentions from both boxes and see one total number.
Combine Both Boxes Cume/Pref Report
Similar to the Rolling Average Cume/Pref report, the Combine Both Boxes report also allows you to roll multiple weeks of data together for each of the weekly cycles. This increases the sample size used to compile each weekly cycle, and will smooth out the peaks and valleys of results from weeks with lower than normal sample sizes.
Cume/Pref Conversion Report
Cume/Pref Conversion Report
The Cume/Pref Conversion report applies a simple formula to the listening responses you interviewers have collected. This report shows you how efficiently each radio station is converting the people that said they listened to each radio station (their responses in the first listening box in Make Calls), to people that said they preferred that station (their response in the second listening box in Make Calls).
The formula itself is actually that simple; it is:
(Pref / Cume) * 100
We divide all the people that said they preferred each station, by all the people that said they listened to that station. As we end up with a fraction (or decimal) answer, such as .412 for KKBT for the week of 4/4 in the example report above, we then multiply the result by 100, to produce a percentage; 41.2% in this case.
If you are using Fave 1/ Fave 2 logic and not collecting all stations cumed, you should not run this report.
In this case, using first preference for the first listening box, and second preference for the second box, will not produce any meaningful results for you with this report. As you don't have an "all stations cumed" variable to divide into the preference number, you will not have the information necessary to determine cume to preference conversion.