The generated variables HGHHSIZE, HGHHAGE01, HGHHAGE02 and HGHHAGE03 were generated simultaneously, with a focus on harmonization. The first step was to aggregate the AP information from the HH matrix and check those and the direct information on HH size and age groups for consistency. If there were large inconsistencies in the information of the APs and more than one interview was available, the direct information on HH size and age groups of the HMs was checked for consistency and used. If both the APs' and HMs' data were highly inconsistent or missing, the HH size of HHs with more than one invited person was set to the number of invited persons (and all to age group 18+), while HHs in which only the AP was invited were set to missing. Source variables: PHHSIZE, PHHAGE01, PHHAGE02, PHHAGE03, HPNAME01-HPNAME12, HPBMNTH01-HPBMNTH12, HPBYEAR01-HPBYEAR12, PCOHAB and PID. Inconsistencies are shown in the variable HGHHSIZE_FLAG. The variables wer filled step by step:
-
AP: include fully consistent cases
-
AP: include slighty inconsistent cases if HH sizes (matrix & response) & generated age groups are consistent
-
AP: include cases if HH size response is missing, but matrix information is consistent
-
AP: include cases if HH size response is smaller than the generated one, but whose matrix information is consistent
-
AP: include cases if HH size & age group responses is are consistent
-
AP: include information with partial missings or inconsistencies if only 1 interview & the reported HH size was greater than 1
-
HMs: include fully consistent cases (intra-person & inter-person)
-
HMs: include cases if the (intra-person & inter-person) information of at least 2 cases is consistent
-
HMs: include cases if the (intra-person) information is consistent, starting with the oldest
-
PBRUTTO: assign no. of invited persons if more than 1 invited person in HH
-
PBRUTTO: assign -1 if 1 invited person in HH
Die generierten Variablen HGHHSIZE, HGHHAGE01, HGHHAGE02 und HGHHAGE03 wurden gemeinsam erzeugt, wobei das Augenmerk darauf lag, dass sie harmonisieren. Dazu wurde in einem ersten Schritt die AP-Informationen aus HH-Matrix aggregiert und diese sowie die direkten Angaben zu HH-Größe und Altersgruppen auf Konsistenz überprüft. Falls bei APs sehr große Inkonsistenzen auftraten und mehr als ein Interview vorlag, wurden die direkten Angaben zu HH-Größe und Altersgruppen der HMs auf Konsistenz geprüft und ggf. verwendet. Falls sowohl die Angaben der APs als auch der HMs hoch inkonsistent waren oder fehlten, wurde die HH-Größe von HH mit mehr als einer eingeladenen Person auf die Anzahl der eingeladenen Personen (und alle auf Altersgruppe 18+) gesetzt, während HH, in denen nur die AP eingeladen wurde, auf missing gesetzt wurde. Quellvariablen: PHHSIZE, PHHAGE01, PHHAGE02, PHHAGE03, HPNAME01-HPNAME12, HPBMNTH01-HPBMNTH12, HPBYEAR01-HPBYEAR12, PCOHAB und PID. Inkonsistenzen werden in der Variable HGHHSIZE_FLAG ausgewiesen. Der Befüllung der Variablen erfolgte schrittweise:
-
AP: include fully consistent cases
-
AP: include slighty inconsistent cases if HH sizes (matrix & response) & generated age groups are consistent
-
AP: include cases if HH size response is missing, but matrix information is consistent
-
AP: include cases if HH size response is smaller than the generated one, but whose matrix information is consistent
-
AP: include cases if HH size & age group responses is are consistent
-
AP: include information with partial missings or inconsistencies if only 1 interview & the reported HH size was greater than 1
-
HMs: include fully consistent cases (intra-person & inter-person)
-
HMs: include cases if the (intra-person & inter-person) information of at least 2 cases is consistent
-
HMs: include cases if the (intra-person) information is consistent, starting with the oldest
-
PBRUTTO: assign no. of invited persons if more than 1 invited person in HH
-
PBRUTTO: assign -1 if 1 invited person in HH