Defaults/Configure tab settings
The Completes Only switch specifies whether only scores from respondents who completed the test will be used in any reports you run. A complete is defined as someone who gave a score for all the songs you tested for the test cycle in which they completed the test.
When Completes Only is on, scores from incomplete test takers (Take Test/X status) are not deleted from your system, but they're simply not included in the sample for any reports you run. You can change this setting, run a report(s), and then change it back again to see what effect including non-completes scores in your sample will have.
Remember, as each respondent begins at a random point in your song test list, the scores missing for any songs by incompletes will vary; it won't simply be missing scores for the last 5 songs in your test, for example. To determine whether to include all scores from all respondents (Completes Only unchecked), or to include scores from only completes, you should run a Scores Integrity Report for the current test cycle. This report will tell you how many scores you have from each respondent that didn't complete the test. In addition, you can run Delete Incompletes, and set a threshold. For example, you might wish to delete all scores from respondents who didn't complete at least 10 songs in your test, but leave all scores for any respondent who did score at least 10 songs.
Weighting will weight up or down the scores for any report based on demographics (age and gender), ethnicity and station preference. The amount of weighting is determined by how closely the actual respondent's profile matches the weighting you defined in the Set Recruitment Quotas screen. For example, if you specified that you wish to test 70% Caucasian and 30% Hispanic, the results will be weighted accordingly. This isn't a big problem if the actual test results are close to this ideal; for example if they're actually 65/35. But if the results are vastly different, for example 40/60, then the amount of weighting that must be applied to achieve your desired result will cause the scores to vary wildly from the actual results. You may wish to run an Incidence Report first, and compare how closely your in-tab respondents match your Quotas. You can also run the same report twice; once with Weighting turned on and again with Weighting turned off to see what effect it has on the results.
Scores Per Song Threshold
This setting allows you to dictate how many scores a song must have for a given research cycle, before it begins showing up in reports. The practical application of this is that you will not see any results for any songs, until that song has at least the number of scores you have indicated in this field. Thus, if you put a new song in on Thursday night, and you run results Friday morning, you won't see results for that new song mixed in with all the other existing songs, if it hasn't yet reached your Threshold setting.
To determine the correct setting for this field, run a Raw Counts report to the screen.
While most reports in ComQuest indicate an aggregate sample size at the top of the report, Raw Counts is the only report which will display the exact number of scores collected for each song. For example, if you run Raw Counts for the last week, most songs might have 90 scores beside them, but there are a few that only have 10 scores. If you wish to filter these songs out of all of your reports, then you would enter 11 (or more) in the Threshold field. Since 10 is less than 11, all the songs with 10 scores will be filtered out (not displayed), while the songs with 90 scores will be included in each report and displayed. In the screen shot above, we have set the Scores Per Song Threshold to 12, meaning any song (for whatever date or demo range we run) that has fewer than 12 scores, will be suppressed, and not ranked or displayed in any reports.
Gender Mix is designed to alternate 50/50 between Males and Females in Make Calls. If most of your database has already been screened and the gender field is completed for most respondent records, this feature will work pretty well for you at attaining a 50/50 mix of the sexes. If a majority of your database is one gender or the other, the feature will work pretty well in the beginning, but eventually you will run out of qualified names for the minority gender, and the system will be substantially slowed down as it attempts to find a new record from the minority gender.
If the gender is not indicated on a majority of records in your respondent file this feature will not work for you, and in fact can cause your system to freeze or slow down.
Gender Mix is designed for use by stations that already have a gender-coded database, comprised of almost equal Male & Female respondents, and that conduct callout for both Men and Women (50/50) each research cycle.
This setting is for use only by radio stations that meet the following criteria:
|•||Make a majority of their calls during the day (9am - 5pm);|
|•||Have different daytime phone numbers (stored in the Daytime field) for these respondents.|
If you don't meet both of these criteria, you should leave this field set to 00:01. If you do wish to use this feature, here's how it's used.
If you make a vast majority of your calls before 5pm, and you have daytime (or work) phone numbers for many of these respondents, you would set this field to 5:00pm. This tells ComQuest that before this time (5:00pm), you wish it to dial (only) respondents at their Daytime number. This is a separate number that is stored in their respondent record. Only people who have a phone number in this Daytime field will be considered for calling before 5:00pm. Then, after 5:00pm (if that's the time you have set in this field), all respondents will be called using their primary respondent phone number (not their Daytime number).
Again, unless you meet both of the criteria above, you should not use this feature (leave it set to the default of 00:01).
When a Busy signal or Answering Machine is reached after dialing, the interviewer will select that button on the screen. The respondent record will be tagged with a recall time, based on what you have entered in the next two fields. Generally, a Busy signal indicates that there is someone at the residence, but they are currently on the phone. A suggested recall time would be somewhere between 10-15 minutes from now. You want to try them back, hopefully when they're off the call they're now on, but before they leave the house or go do something else.
An Answering Machine implies the person is not there, or they are screening their calls. In this case, a longer Call Back is recommended; perhaps 55-65 minutes. With about a one-hour call back for Answering Machines, it's conceivable that some people that are not at home might get called back two or three times in an evening. This is not bad, although it does slightly lower the productivity of your interviewers.
This is simply a line of text that will appear at the top of all your printed reports. It's especially helpful if you are conducting callout for two or more stations from your facility. Each database can have their own unique Report Header, so you will easily be able to ascribe Incidence and Cume/Pref reports, etc. to the correct station, based on this Report Header.
Don't Use Panels In Cume/Pref Reports
This is a setting that was added to filter out panels (previous test-takers) from all Cume and Preference reports. The logic is that as these people have previously been included in a Cume and Pref sample (when they were first recruited), they should not continue to be counted when we subsequently call them back for more song testing. Including them in an otherwise random sampling of respondents would tend to inflate the Cume and Pref results for any stations that are part of your recruitment quotas for the song test.
The setting for this variable will eventually be moved to the Override Settings screen.
If you are conducting callout for two or more stations in the same market, you will probably be exporting respondent records from one station's database to another from time to time. This is where you set the default path for the subdirectory to which you will export respondent records from this database. The form for the path is:
Generally, this would be:
or whatever the subdirectory name for where the other station's data is located. If you don't have another station you're exporting to, you should leave this set to the default: C:\RESPONDENTS. This way, in case you inadvertently go in to the Export Records procedure, you won't get a fatal error (which is what happens if you have no path for exporting.)
When you are done entering your defaults, be sure to click on OK (or press Enter) to save your work. Selecting Cancel will exit the Default screen without saving any changes.