FIELDFOLLOWS contains one row for each package of Mother-Infant follow data put together by any one field researcher during the course of a day. There is one field follow for every researcher for every mother for each day for which the researcher collected data on the mother (and her infant). The mother (or infant) need not be under continuous observation for the interrupted observations to be part of a single field follow, in fact each row in FIELDFOLLOWS represents an entire day's worth of a researcher's observations of a mother (and infant). All the field follows, irrespective of the researcher in the field, which record data on any one mother (and her offspring) in any one day are combined for analysis into an “analytic follow”, usually referred to herein just as a “follow”, and are represented in the database as rows in the FOLLOWS table.
Each row in FIELDFOLLOWS must have at least one related row on Sheets.
An observer cannot be in two places at the same time -- the combination of Collector, FOLLOWS.Date, and INTERVALS.Time must be unique.
The date of the follow, the FOLLOWS.Date value, cannot be after the date of data entry, the Entrydate value. Likewise, the date of data entry, the Entrydate value, cannot be after the date the data is uploaded into the Gombe-MI system, the UploadDate value.
The date of data entry,
Entrydate, may not be NULL when
the FOLLOWS.Date is
after 2011-06-30.
The GCStatuscolumn must be
non-NULL when there is any group composition information
recorded in the database which is related to the field follow,
otherwise it must be NULL. This is tested for by checking for
the existence of rows in the RAW_GROUPS table
related by SheetID to the
Sheets table, related by
FFID to the FIELDFOLLOWS table. This
check is performed upon transaction commit.
A unique integer identifying the field follow. This column is automatically maintained by the
database, cannot be changed, and must not be
NULL.
Code identifying the person who observed the follow and recorded it's information on the template sheets. The legal values for this column are defined by the PEOPLE support table.
This column may not be NULL.
The date on which the field follow's data entry was finished.
This column may sometimes be NULL, see above.
Code identifying the person who entered the field follow information recorded on the template sheets into Gombe-MI. The legal values for this column are defined by the PEOPLE support table.
This column may not be NULL.
The date on which the field follow's data was uploaded into the Gombe-MI system.
When values for this column are not supplied as part of
the upload process, or a NULL is supplied as a data value, the
UploadDate value defaults to the current date. This column may not be NULL.
FOLLOWPARTS contains one row for every Mother-Infant follow for every individual in the family that is followed. There are at least 2 participants in every follow: there must be exactly one row for the mother and at least one row for an infant. These rules are enforced on transaction commit.
Each participant in a follow must be a member of the study population on the date of the follow. The BIOGRAPHY_DATA.Entrydate cannot be after the follow's FOLLOWS.Date value. The BIOGRAPHY_DATA.Departdate cannot be before the follow's FOLLOWS.Date.
The mother in a follow must be the mother of the infant (either biological or an adoptive mother) and, similarly the mother of, if any, the sibling -- the BIOGRAPHY_DATA.MomID of the offspring, or one of the offspring's related ADOPTIONS.AParent values, must be the BIOGRAPHY_DATA.AnimID of the mother, where the role of mother and infant or sibling is determined based on FOLLOWROLES.Function value and the AnimID value is determined using the CHIMPIDS table. When the mother is an an adoptive parent then the date of the follow cannot be before the adoption date -- the FOLLOWS.Date value cannot be before the ADOPTIONS.ADate value. The sibling, if any, in a follow must be older than the infant. Two participants in a follow may have the same role[12] (e.g. twin) if and only if the two participants have the same birth date.[13]
When twins are both participants in one follow the two twins must share a role in the follow -- must have identical Role values.
The individual(s) in the follow with the role of infant or sibling must be less than 15 years old on the date of the follow.
There are no checks to ensure that the infant in the follow is the youngest living offspring, or that the sibling is the next youngest.
An individual cannot participate twice in a single follow -- there cannot be 2 FOLLOWPARTS rows with the same FollowID and related AnimID values.
A unique integer identifying the participant in the
follow. This column is automatically maintained by the
database, cannot be changed, and must not be
NULL.
The integer identifying the follow of the participant. The FollowID value cannot be changed, the row must be deleted and a new row inserted in it's place.
This column may not be NULL.
The identifier assigned to the follow participant. The ChimpID code of the individual.
This column may not be NULL.
Code indicating the individual's role in the follow. The legal values for this column are defined by the FOLLOWROLES support table.
This column may not be NULL.
FOLLOWS contains one row for every Mother-Infant follow. There is one follow for every mother for each day for which there is a record; a chimpanzee cannot be the focal in more than one follow on any given day. The mother (or infant) need not be under continuous observation for the interrupted observations to be part of a single follow, nor do all observations need to be made by the same person.
Each follow must have at least one time interval during which observation was made -- each FOLLOWS row must be at least one related row on INTERVALS.[14] This check is performed upon transaction commit.
Deleting a row from the FOLLOWS table will delete the entire follow -- the system will first delete all related INTERVALS and FOLLOWPARTS rows, which will in turn delete dependent rows from other tables.
A unique integer identifying the follow. This column is automatically maintained by the
database, cannot be changed, and must not be
NULL.
INTERVALS contains one row for each interval of time during which observations were recorded as part of a follow.
At the time of this writing all follow observation intervals are of 1 minute duration. The Gombe-MI system contains elements which rely on per-minute granularity.[15]
The time intervals recorded on any given template sheet must be part of a single follow -- all of the INTERVALS.FollowID values must be identical for those INTERVALS rows having a given SheetID value.
A follow cannot have overlapping intervals of observation -- the combination of Time and FollowID must be unique.
All of the time intervals related to a given group composition survey must be part of a single follow -- all of the INTERVALS.FollowID values must be identical for those INTERVALS rows having a given GID value.
Deleting a row from the INTERVALS table will delete entire interval -- the system will first delete all related DISTANCES, EVENTS, and Sheets rows, which will in turn delete dependent rows from other tables.
A unique integer identifying an interval of observation
taken during a particular follow. This column is automatically maintained by the
database, cannot be changed, and must not be
NULL.
The integer identifying the follow to which the observation interval belongs.
This column may not be NULL.
The integer identifying the template sheet on which the interval observation was recorded.
The SheetID value cannot be changed. This column may not be NULL.
The time at the start of the interval of observation. The
stored value has a one minute precision; the seconds portion of
the Time value must be 0.
The Time value cannot be
before 06:00 and cannot be
after 20:00. This column may not be NULL.
The unique identifier assigned to each record of group composition. The GROUPS.GID code of the group composition time period into which the observation interval falls.
Specifying a value of NULL causes the system to
automatically compute this value.
The GID value may be NULL if the observation interval is
not associated with a period of time during which group
composition was recorded.
Textual comments concerning the entire observational interval. These comments are those of the data collector in the field.
This column may not be empty, it must contain characters,
and it must contain at least one non-space character. However, the Comment value may be NULL when
there are no comments concerning the entire observation
interval.
SHEETS contains one row for every template sheet used in the field when recording follows, or in those cases where there is no longer an electronic record of exactly which intervals of a follow were written on which physical piece of paper, one row per follow per field observer involved in recording the follow (i.e. per FIELDFOLLOWS row.
Every template sheet must have a record of at least one interval -- each SHEETS row must have at least one related INTERVALS row. On the SHEETS table this condition is checked at transaction commit time, although on the INTERVALS table the check is done immediately in the case of UPDATE and at transaction commit time in the case of DELETE.
Use transactions when making changes to both SHEETS and INTERVALS to produce bundles that ensure that at transaction commit each row on SHEETS has at least one related row on INTERVALS. Insert rows first into SHEETS and then into INTERVALS, and reverse this order when deleting rows.
Deleting a FIELDFOLLOWS last related row from the SHEETS table will delete the FIELDFOLLOWS row.
The Page value must be unique per
field follow (per FIELDFOLLOWS row) and there
may be at most one NULL Page value per
field follow.
A unique integer identifying the sheet. This column is automatically maintained by the
database, cannot be changed, and must not be
NULL.
The integer identifying the field follow, the set of sheets collected by a researcher in the field comprising a day's mother-infant follow, in which the sheet belongs.
The SheetID value cannot be changed. This column may not be NULL.
Code identifying the type of the follow; the type of data entry template used for data collection that determined what was on the printed page.[16] The legal values for this column are defined by the FOLLOWTYPES support table.
This column may not be NULL.
The page number written on the sheet. This number must be an integer larger than zero. Note that the system does not require the page number to be contiguous.
This value may be NULL when the page number is not known
or not recorded in electronic form.
The name of the valley or other physical location where the data written on the sheet was recorded.
This column may not be empty, it must contain characters,
and it must contain at least one non-space character. However, this column may be NULL when there
is no recorded location.
[12] Technically, may have the same FOLLOWROLES.Function value.
[13] Or if any of the birth dates involved are NULL.
[14] Note however that nothing need be observed during a follow. Although the INTERVALS row must exist there is no requirement that an INTERVALS row have any related data, so the system does not require any behavioral or inter-focal distance data.
[15] See Appendix C for a list of the assumptions regarding interval duration.
[16] Presumably, the meaning of the data collected as determined by the data collection protocol in use did not change without a change in the print format of the data collection sheet.