@@ -56,8 +56,40 @@ Beginning in Wave 11, the `module` column indicates which module the respondent
5656was randomly assigned to. See the [ survey modules and randomization] ( modules.md )
5757information for more details.
5858
59- Coding details for each survey wave follow.
59+ ### Privacy Restrictions
6060
61+ To prevent respondents from being identifiable in the response data, responses
62+ with ZIP codes with populations of 100 or fewer have their location set to ` NA ` .
63+ This affects item A3 in the individual response files. (This change was
64+ implemented with the introduction of Wave 4. Previously, all ZIPs were
65+ reported.)
66+
67+ Invalid ZIP codes are preserved unchanged, and these rows are reported in the
68+ individual response files with their invalid ZIPs.
69+
70+ ### Race and Ethnicity
71+
72+ Beginning in wave 4, items D6 and D7 ask respondents for race and ethnicity.
73+ These columns are ** not** available in the microdata files due to
74+ reidentification concerns. Public contingency tables that aggregate by race and
75+ other demographic variables [ are available] ( contingency-tables.md ) .
76+
77+ Users with a specific need for these variables in microdata should contact us at
78+ < [email protected] > to discuss options for obtaining them,
79+ as access can be provided under some restrictions.
80+
81+ In the contingency tables, and in the microdata files for users who have gained
82+ permission to access this data, the ` raceethnicity ` column is coded based on the
83+ following rules:
84+
85+ * If the respondent answers "Yes" to item D6, they are coded as Hispanic,
86+ regardless of their answer to D7. If the respondent answers "No" to D6, the
87+ following rules apply.
88+ * Respondents who selected more than one racial group in D7, or who selected
89+ "Some other race", are coded as "NonHispanicMultipleOther".
90+ * Respondents who selected only one racial group in D7 are coded according to
91+ that race, such as "NonHispanicAsian" or
92+ "NonHispanicAmericanIndianAlaskaNative".
6193
6294## Comprehensive Codebook
6395
@@ -91,23 +123,6 @@ Available columns:
91123* ` group_of_respondents_item_was_shown_to ` : Module-based subset of respondents item was asked of. One of "all" (if item was included in the [ Daily Core] ( modules.md ) and asked of all respondents), "Module A", or "Module B".
92124
93125
94- ## Privacy Restrictions
95-
96- To prevent respondents from being identifiable in the response data, responses
97- with ZIP codes with populations of 100 or fewer have their location set to ` NA ` .
98- This affects item A3 in the individual response files. (This change was
99- implemented with the introduction of Wave 4. Previously, all ZIPs were
100- reported.)
101-
102- Invalid ZIP codes are preserved unchanged, and these rows are reported in the
103- individual response files with their invalid ZIPs.
104-
105- Additionally, beginning in wave 4, items D6 and D7 ask respondents for race and
106- ethnicity. These columns are ** not** available in the microdata files due to
107- reidentification concerns. Users with a specific need for these variables should
108- contact us at
< [email protected] > to discuss options for
109- obtaining them. However, contingency tables that aggregate by race and other
110- demographic variables [ are available] ( contingency-tables.md ) .
111126
112127
113128## Wave 1
@@ -471,7 +486,7 @@ new items were meant to capture reasons for vaccine hesitancy among respondents.
471486
472487### New Items
473488
474- * Item V2a ask respondents that have received a COVID-19 vaccine and indicated
489+ * Item V2a asks respondents that have received a COVID-19 vaccine and indicated
475490 that they have not had 2 doses of the vaccine whether they intend to get the
476491 required doses.
477492* Items V5a-V5d and V6 were added to capture reasons for respondents not
@@ -775,6 +790,11 @@ are not anticipated to change the respondent’s answer.
775790 might become seriously ill from CVOID-19?”) was removed.
776791* Item C17a (“Have you had a seasonal flu vaccination since July 1, 2020?”) was
777792 removed.
793+ * On November 8, 2021, item V2 “How many vaccinations have you received?” was
794+ removed from Wave 11, as the response options did not allow for respondents to
795+ report booster shots and additional doses. Booster doses were becoming more
796+ common at this time, making the lack of booster response options jarring to
797+ respondents.
778798
779799### Notes
780800
@@ -789,36 +809,43 @@ are not anticipated to change the respondent’s answer.
789809
790810## Wave 12
791811
792- Wave 12 was deployed on October 7, 2021. Deployment is in two phases:
812+ Wave 12 was deployed on December 19, 2021, after an experimental phase in
813+ October:
793814
794- * Initially the survey was given only to 15% of survey respondents, the
795- remainder receiving Wave 11. This phase lasted until October 22. This phase
796- was used to conduct three experiments with (1) demographic module placement,
797- (2) vaccination uptake question, and (3) survey invitation text. Details of
798- these experiments can be found [ below] ( #experiments ) . In data files, this is
799- marked as wave "12.5" to distinguish from the final version.
800- * After results of the experiments are analyzed, a final version will be
801- deployed using only one demographic module placement and one vaccine uptake
802- question. This will be marked as wave 12 in the data files.
815+ * From October 7 to October 22, 2021, an experimental Wave 12 was given only to
816+ 15% of survey respondents, the remainder receiving Wave 11. This phase was
817+ used to conduct three experiments with (1) demographic module placement, (2)
818+ vaccination uptake question, and (3) survey invitation text. Details of these
819+ experiments can be found [ below] ( #experiments ) . In data files, this is marked
820+ as wave "12.5" to distinguish from the final version. Data from the
821+ experimental wave 12 was not included in the aggregates published in the
822+ COVIDcast API or in our [ contingency tables] ( contingency-tables.md ) .
823+ * After results of the experiments were analyzed, the final version was deployed
824+ on December 19, 2021. In data files, this is marked as wave 12.
803825
804826Wave 12 is available in English and the languages introduced in Wave 3.
805827
806828Files:
807829* [ Wave 12 text and coding] (waves/CMU Survey Wave 12.pdf) (PDF)
808830* [ Wave 12 text and coding] (waves/CMU Survey Wave 12.docx) (Word)
809831
810- Besides the experiments, Wave 12 revises the schooling module of the survey
832+ Besides the experiments, Wave 12 revised the schooling module of the survey
811833instrument and adds questions regarding parents’ intention to vaccinate their
812- children (under age 18). Please review the changes carefully when you use
834+ children (under age 18). These schooling items are part of [ Module B] ( modules.md )
835+ of the survey. Please review the changes carefully when you use
813836responses from multiple waves of this survey.
814837
838+ ### Demographic Module Placement
839+
840+ In the prior waves of CTIS, respondent demographics were collected towards the
841+ end of the survey. These questions are now asked after the vaccine module of the
842+ survey. This placement was chosen after the [ order experiments described
843+ below] ( #experiments ) found that this ordering improved response rates to the
844+ demographic items. We expect this will aid users who need demographic data for
845+ their survey analyses.
846+
815847### New Items
816848
817- * Item V1alt asks respondents who they personally know has already received a
818- COVID-19 vaccine. Response options include themselves, household members, and
819- others. This is an alternate form of the item capturing vaccine uptake, and is
820- part of the vaccination uptake question experiment described
821- [ below] ( #experiments ) .
822849* Item P1 asks if the respondent is a parent or legal guardian of a child under
823850 age 18.
824851* Item P2 asks respondents the age group of their oldest child. This replaces
@@ -832,16 +859,18 @@ responses from multiple waves of this survey.
832859* Item P6 asks respondent about the preventative measures that apply if their
833860 oldest child is attending any in-person classes. These response options are a
834861 subset of response options from E3 with several new additions.
862+ * In the experimental phase, item V1alt asked respondents who they personally
863+ know has already received a COVID-19 vaccine. Response options include
864+ themselves, household members, and others. This is an alternate form of item
865+ V1 capturing vaccine uptake, and is part of the vaccination uptake question
866+ experiment described [ below] ( #experiments ) . In the final version of Wave 12,
867+ item V1alt is not used and item V1 remains unchanged.
835868
836869### Changed Items
837870
838871* Item B13a asks respondents “Have you ever had coronavirus (COVID-19)?” and
839872 replaces item B13 “As far as you know, have you ever had coronavirus
840873 (COVID-19)?” for ease of translation.
841- * The vaccination uptake question experiment uses questions V1 and V1alt.
842- Subsequently, questions that use V1 in the display logic have been updated to
843- also include respondents who are asked V1alt. The display logic has been
844- updated for the following questions: V2, V11a, V3a, V5a, V12a, V16, and V9
845874* Item V5a, V5b and V5c have the following revisions to response options:
846875 * The response option “I don’t like vaccines” was revised to “I don’t like
847876 vaccines generally” to distinguish from a specific dislike of COVID-19
@@ -855,6 +884,11 @@ responses from multiple waves of this survey.
855884 added: “The available appointment locations did not work for me” and “Other”
856885* Item V15b has two new response options: “The available appointment locations
857886 did not work for me” and “Other”.
887+ * The vaccination uptake question experiment, part of the experimental version
888+ of Wave 12, used questions V1 and V1alt. During the experiment, questions that
889+ use V1 in the display logic were updated to also include respondents who are
890+ asked V1alt. The display logic was updated for the following questions: V2,
891+ V11a, V3a, V5a, V12a, V16, and V9.
858892
859893### Removed Items
860894
@@ -871,41 +905,45 @@ responses from multiple waves of this survey.
871905
872906Wave 12 was initially launched to 15% of respondents to collect data for the
873907experiments described below. (The remaining 85% continued to receive Wave 11.)
874- After the experiments have been conducted and the data analyzed; the CTIS team
875- will finalize the Wave 12 instrument to be distributed to the entire sample.
876- Wave 12 will consist of three experiments with (1) demographic module placement,
877- (2) vaccination uptake question, and (3) survey invitation text.
908+ This experiment was run from October 7 to October 22, 2021. After the
909+ experiments were conducted and the data analyzed, the CTIS team finalized the
910+ Wave 12 instrument to be distributed to the entire sample. This finalized
911+ version was deployed on December 19, 2021.
912+
913+ The Wave 12 trial phase consisted of three experiments with (1) demographic
914+ module placement, (2) vaccination uptake question, and (3) survey invitation
915+ text.
878916
879917#### Demographic Module Placement and Vaccine Uptake Question
880918
881- The 15% of respondents who receive the experimental Wave 12 will be divided into
919+ The 15% of respondents who received the experimental Wave 12 were divided into
882920three groups:
883921
884- * Group 1 (5% of respondents) will see the demographics questions after the
885- symptom module, which concludes with question B2b “For how many days have you
886- had at least one new or unusual symptom?” This group of respondents will see
887- the Wave 11 version of the vaccine uptake question (V1).
888- * Group 2 (5% of respondents) will see the demographics questions after the
889- COVID vaccine module, which concludes with question V9 “How concerned are you
890- that you would experience a side effect from a COVID-19 vaccination?” This
891- group of respondents will see the Wave 11 version of the vaccine uptake
892- question (V1).
893- * Group 3 (5% of respondents) will see the experimental form of the vaccination
922+ * Group 1 (5% of respondents) saw the demographics questions after the symptom
923+ module, which concludes with question B2b “For how many days have you had at
924+ least one new or unusual symptom?” This group of respondents saw the Wave 11
925+ version of the vaccine uptake question (V1).
926+ * Group 2 (5% of respondents) saw the demographics questions after the COVID
927+ vaccine module, which concludes with question V9 “How concerned are you that
928+ you would experience a side effect from a COVID-19 vaccination?” This group of
929+ respondents saw the Wave 11 version of the vaccine uptake question (V1).
930+ * Group 3 (5% of respondents) saw the experimental form of the vaccination
894931 uptake question (V1alt), which asks respondents “Do you personally know anyone
895932 who has received the COVID-19 vaccine already?” Response options include
896- themselves, household members, and others. These respondents will not receive
897- the previous vaccine uptake question, V1. They will see the demographic
898- questions in the Wave 11 placement, at the end of Module A/B and before the
899- occupation questions.
933+ themselves, household members, and others. These respondents did not receive
934+ the previous vaccine uptake question, V1. They saw the demographic questions
935+ in the Wave 11 placement, at the end of Module A/B and before the occupation
936+ questions.
900937
901- A new column ` w12_treatment ` in the microdata files will indicate which of the
902- three groups above each respondent was assigned to.
938+ A new column ` w12_treatment ` in the microdata files indicates which of the three
939+ groups above each respondent was assigned to.
903940
904941#### Survey Invitation Text
905942
906- Our partners at Facebook will also be conducting an experiment among those
907- allocated to receive Wave 11. Among those participants, 6% will receive will
908- receive one of six different invitations to the survey on their News Feed. This
909- will be an internal Facebook experiment conducted with the goal of improving
910- response rates and exploring additional non-response bias in the weighting
911- process. More details will be provided when the survey invitation is finalized.
943+ During the Wave 12 experiment period, our partners at Facebook conducted an
944+ experiment among those respondents allocated to receive Wave 11. Among those
945+ participants, 6% will receive will receive one of six different invitations to
946+ the survey on their News Feed. This will be an internal Facebook experiment
947+ conducted with the goal of improving response rates and exploring additional
948+ non-response bias in the weighting process. More details will be provided when
949+ the survey invitation is finalized.
0 commit comments