Audio audit

Audio auditing enables audio recording of the interviewing process for quality control.

Survey Solutions offers two modes for audio audit:

full: the whole interview is recorded.
selective: some parts of an interview are recorded (as decided by the HQ staff).

It is possible to:

specify which assignments will be recorded and which will not.
specify which mode (full or selective) is applicable to which assignment.
define the scope of the selective audit recording differently for different assignments.

Mobile devices store the recording in encrypted audio files, and synchronize these audio files along with the corresponding interview data to the server allowing for subsequent quality control.

How to enable/disable full audio auditing

Before creating an assignment

In the main menu of the Headquarters navigate to Survey Setup > Questionnaires;
Click on a questionnaire;
In the context menu that appears, select Details;
Click in the box next to Record audio to enable audio auditing.

After this point, all newly created assignments will have audio audit enabled. Note that the assignments that have been created to this point will retain their assigned audit mode.

To disable audio auditing, follow the same process to deselect the checkbox and switch off audio auditing for all future assignments. Any assignments that have been created until that moment will retain their audio audit recording setting.

For assignments that are being created with a batch upload, the corresponding attribute is _record_audio:

set the value to: 1 to indicate that the audio audit recording should be performed for this particular assignment, or
set it to 0 or leave blank (default) to indicate that no audio audit should be performed for that assignment.

After creating an assignment

In the main menu of the Headquarters navigate to Survey Setup → Assignments;
Click on the value in the AUDIO RECORDING column
In the dialog that appears, check (or uncheck) the checkbox Enable audio recording to enable (disable) audio auditing for selected assignment.

This action affects only the selected assignment (even if multiple assignments are selected in the assignments list).

How to enable/disable selective audio auditing

Selective audio audit is based on a scope defined by the user creating an assignment. This scope identifies which parts of the interview must be recorded, by mentioning the relevant identifiers as a comma-delimited list in the _aascope property during the batch upload of assignments, for example: demographics, health, labor, HHMEMBERS.

The identifier of the following components may be mentioned:

variable names of sections;
variable names of sub-sections;
variable names of rosters.

Note that the variable names of sections and sub-sections are optional in Survey Solutions. Thus the author of the questionnaire must assign them at the questionnaire design to permit their subsequent audio recording (full audio audit does not have this requirement). To see which identifier was assigned to a certain container the headquarters user may inspect the questionnaire preview document. It is accessible via: Main menu → Survey Setup → Questionnaires → find and click the questionnaire title → Details → click the link corresponding to Questionnaire preview.

Selective audio audit is not applicable directly to rosters in flat display mode. In order to record the audio audit for questions included in such rosters the identifier of the containing section (sub-section, roster) must be mentioned.

Selective audio audit is not applicable to individual questions. To make sure the answer to a particular question is recorded, mention the name of the container for that question (specific section, sub-section, roster).

Selective audio audit is not hierarchical: Survey Solutions Interviewer App will record when the user is in the container mentioned in the scope, but will not record the sub-sections/rosters of that container. If recording of the child containers is desirable, their names should also be mentioned in the scope.

It is not possible to switch off (disable) the selective audio audit or change the scope of the containers to be recorded after an assignment was created, except that the full audit takes precedence over the selective one: if an assignment was created with a selective audit (non-blank scope) an HQ or admin user can subsequently turn on full audio audit for that assignment to record interviews fully. This change will affect only new interviews created from that assignment (after synchronization) and will not affect the interviews that have been already started. A reverse change (switching off the full audio audit for an assignment) will cause it to fall back to the selective audio audit (if the corresponding scope was defined).

On a server

For every assignment in the assignments list there is an indication in the AUDIO RECORDING column regarding the audio audit that will be applied to the interviews originating from this assignment:

Yes - full audio audit will be applied;
Selective - selective audio audit will be applied;
No - no audio audit will be applied.

If this mode has been changed, the mobile devices will need to be synchronized to become aware of this change.

On a mobile device

The Android OS prompts the user to give the Interviewer App the permission to access the microphone. Interviewers will not be able to start interviews that require audio audit without enabling required Android permissions.

If the permission is given and subsequently withdrawn, the interviewer will not be able to continue the interview (until the permission to access the microphone is reissued to the Interviewer App).

Warning:

Modern versions of Android OS indicate to the users when the recording is in progress, thus the interviewers are aware of it.

There is no setting or configuration available to interviewers to affect whether the device will or will not record the audio for the purpose of audit or the quality of such an audio recording.

Important considerations

Audio recordings will increase the amount of data that needs to be synchronized with the server. The amount of audio audit data depends on the quality setting and can be estimated using the table contained in Format of Audio Files.

Audio audit occurs only on a mobile device (does not apply to web interviews) and only when the device is on and the interview is in progress (but not while the interviewer is reviewing the dashboard, working with the settings, etc).

For technical reasons, the audio recording for audit purposes is not performed during the following situations:

answering of a geography question;
selecting a file from the gallery;
taking a photograph with the camera;
scanning a barcode/qrcode with the camera;
adjusting the App’s settings, running diagnostics, etc.
answering audio question (nb: recording is taken as the answer to the audio question and not added to audio audit recording).
viewing map data in an external application (following the link shown in GPS questions);
switching to any other App, or
switching to home screen.

For technical reasons, interviews and their audio files must be sent together. The audio files are synced first, and then the interview file. That means the interview file will only be sent once the, potentially quite large, audio files have been sent.

The recording is done using the currently set up microphone, which is typically the device’s built-in microphone, but external microphones can also be used.

Warning:

Due to modern OS security restrictions (to protect the privacy of phone conversations), audio audit is not applicable for the cases when the interviewing is done via a phone call from the same device (i.e. when the Interviewer App is running on the same smartphone used for the call). In such a case the audio audit may not record the remote party (only the interviewer’s voice will be recorded since the device’s microphone isn’t exposed to the respondent) or may be disrupted altogether (depending on the device manufacturer and OS version).

Exporting audio audit recordings

Audio audit data is a separate type of data export (separate from e.g. main survey data, or binary data). Audio audit recordings are exported for interviews passing the export selection (by status, date and time), and are placed in subfolders corresponding to individual interviews.

There can be (and often are) multiple audio audit records for a single interview due to interruptions of the recording caused by the events listed in the Important Consideratins section above, or because of the occurrence of non-recordable sections in the selective audit. The program or person conducting the analysis of the audio audit data should be aware of this and be able to process multiple files in accordance with their timestamps.

See audio audit files for specific details on the naming of the audio audit files (the same pattern is used whether the audio audit is full or partial).

More information

For details on the quality settings for audio files see format of audio files.
The naming of the audio audit files is explained in audio audit files.
App permissions are described in Android App Permissions.