These properties are stored in DPM in Sample Managements>ProjectName>Queueing>CATIParameters. If you do not see one of these properties in DPM, it is because UNICOM Intelligence Interviewer is using the default setting for that property. If you need to change the default you can do so by adding the property to DPM using DPM Explorer and giving it the value you require. However, you should not normally need to do this and it is therefore advisable to discuss your requirements with your support representative before making any changes. Only the PredictAdjust property can be changed at will using the Interviewing Options activity in UNICOM Intelligence Interviewer - Server Admin.

Silent calls occur when connections are achieved but no interviewers are free to take the calls. The TargetPredictiveSilentCalls property is a value in the range 0 to 100 that defines the overall percentage of silent calls that you hope to achieve for the project. The default is 2.5. As dialing takes place, the predictive dialing controller continually adjusts the dialing rate in order to try to reach this percentage of silent calls.

Whether or not the dialing provider dials predictively depends on the current percentage of silent calls. The MaxPredictiveSilentCalls property defines the maximum percentage of silent calls per total connected calls that will be allowed per day, where a day is midnight to midnight . The default is 3.0%. Over-dialing is limited to try to ensure that this percentage is not exceeded, but note that it is still possible to exceed this limit if interviewers put their extensions on-hook before closing the Phone Participants activity.

The EnforceMaxSilentCalls property setting determines whether or not the MaxPredictiveSilentCalls property is strictly enforced. The default value is False. When set to True, the property strictly enforces the silent call limit. This means that when the default predict script determines how many calls to dial, it will overdial only as many calls as can be returned as silent without exceeding the silent call percentage limit (as defined by the MaxPredictiveSilentCalls property).

The silent call percentage is determined by dividing the connected calls by the silent calls. At the start of a shift when you have only a few interviewers on a project (and especially if your participant list is such that most calls are not answered, busy, or anything but connected), it can take quite a while before you over-dial. For example, when the silent call limit is 3%, you must connect (1/x)*100=3 calls (33 calls) before you can over-dial by 1. The regulatory bodies that set silent call limits typically allow the limits to be temporarily exceeded, provided the average over a 24 hour period is under the limit.

The MinPredictWaitingInterviewers property determines the number of interviewers that need to be waiting (the dialer is dialing calls for them) before the default predict script will over-dial. The default setting is 0. A zero setting provides more predictive dialing with fewer interviewers.

The PredictAdjust property is a value in the range 0 to 1000 that determines how aggressive predictive dialing will be. In simple terms, the more aggressive the dialing, the more numbers will be dialed per interviewer. A value of 0 disables predictive dialing (over-dialing) so that only one call is dialed per interviewer. Values in the range 1 to 1000 represent the probability that the dialer will make a silent call. For example, if PredictAdjust is set to 50, and connect rate is 50%, the over-dial rate will be 1.5 times the number of waiting interviewers. The default setting is 150.

The PredictAdjust value represents the initial dialing aggressiveness. Over time, the default algorithm will attempt to tune the aggressiveness to achieve the required percentage of silent calls.

The Predict1PI algorithm uses the value of PredictAdjust as the initial setting for the ControllerIntegralGain property that it uses for adjusting the dialing rate. As predictions are made, the algorithm maintains an internal prediction adjustment value but does not update the PredictAdjust property to match. Changes that supervisors make to the property (using the Dialer Settings tab in Interviewing Options) cause ControllerIntegralGain to be reset to match the new PredictAdjust value.

The value of PredictAdjust indicates your trust in the dialing statistics and the way that the predictive algorithm is using them to dial numbers. For example, if you are getting a lot of silent calls and you think that you could reduce this by dialing only half as many numbers per interviewer as is currently the case, you could set PredictAdjust to 5. On the other hand, if you have interviewers waiting too long for numbers you might set PredictAdjust to 30 to dial three times as aggressively (that is, three times more numbers per interviewer) as is currently the case.

The predictive algorithm uses the values of the ControllerProportionalGain and ControllerIntegralGain properties in the calculations of how many calls the dialer should make in order to achieve a connection for every waiting interviewer. ControllerProportionalGain changes the prediction adjustment value in direct proportion to the difference between the target and actual silent call percentages. It defaults to 2. ControllerIntegralGain changes the prediction adjustment value by a small amount for every call attempt based on the difference between the target and current silent call rates. It defaults to 0.05. You should not need to change these properties.

For a change to the CATI parameter settings to be recognized for a given project, without the need to restart the interviewing engines, the Sync property for the sample management needs to be incremented (Site\Servers\SampleManagements\<project>\Sync). Similarly, for changes to the predictive algorithm to be recognized, the DialingSync property needs to be incremented.

For information on how these properties control predictive dialing, see Predictive dialing algorithms.