A Brief Guide to data sources and variables in the public HCMST, version 4.00

Revision Date: 4/13/2014

See also the other online documentation, including the Codebook. The Codebook contains frequencies of all variables, plus a list of notes attached to each variable; notes will be visible from Stata but may not be visible from other software programs. Notes identify, among other things, which variables were produced by Knowledge Networks (now known as GfK), and which variables were produced by the Stanford Research Team. The variables below are only a small fraction of the total set of the variables in the public dataset, but enough (we hope) to give the user a useful idea of what is in the dataset. Also note: this user’s guide is not designed for printing.

Data Sources	Background Demographic Surveys: Knowledge Networks’ Core Adult Profile and Public Affairs Profiles	How Couples Meet and Stay Together Main Survey	Background Demographic Survey Repeated: Core Adult Profile “pp2”	How Couples Meet and Stay Together First Follow-Up “Wave 2”	Background Demographic Survey Repeated Again: Core Adult Profile “pp3”	How Couples Meet and Stay Together Second Follow-up “Wave 3”	Backround demographic survey repeated again: core adult profile “pp4”	how couples meet and stay together third follow-up “wave 4”
Order Data Was Gathered	1	2	3	4	5	6	7	8
Sampling Frame, Theoretical	English Literate Adults living in the United States	English Literate Adults living in the United States who had either a spouse or a romantic partner at time of main survey	English Literate Adults living in the United States	English Literate Adults living in the United States who had either a spouse or a romantic partner in 2009	English Literate Adults living in the United States	English Literate Adults living in the United States who were still partnered with a spouse or a romantic partner from 2009	English Literate Adults living in the United States	English Literate Adults living in the United States who were still partnered with a spouse or a romantic partner from 2009
Sampling Frame, Actual	Members of Knowledge Networks’ Nationally Representative Survey Panel	Members of Knowledge Networks’ Nationally Representative Survey Panel	The 4,002 original survey respondents (survey firm Knowledge Networks continues to gather data on subjects whether they were partnered or not)	The 3,009 survey respondents who were partnered at the time of the main survey (the HCMST follow-up surveys only follow subjects who were partnered at the time of the main survey in 2009)	The 4,002 original survey respondents	The 3,009 survey respondents who were partnered at the time of the main survey [minus the respondents who reported having broken up with their partner, or partner passed away by Wave 2]	The 4,002 original survey respondents	The 3,009 survey respondents who were partnered at the time of the main survey [minus the respondents who reported having broken up with their partner, or partner passed away by Wave 3]
How many Cases are Covered by This Data Source	4,002	3,009/4,002	3,693/4,002	2,520/3,009	2,644/4,002	1,960/3,009	2,035/4,002	1,536/3,009
Response Rate	4,002/4,002=100% since completion of the background demographic surveys was necessary for subjects to join the KN panel. Earlier parts of the response rate can be estimated: 32.6% of people initially contacted by phone agreed to join the KN panel, and of these 56.8% completed the KN Core Adult Survey, estimates from Callegaro and DiSogra 2008  “Computing Response Metric for Online Panels,” Public Opinion Quarterly 72: 1008-1032	71%	3,693/4,002=92.3%	2,520/2,981=84.5% [denominator excludes 28 cases who were determined to be ineligible based on main survey answers]	2,644/4,002=66%	1,960/2,689=72.9% [denominator excludes previous break-ups and other ineligibles]	2,035/4,002=50.9%	1,536/2,557=60.0% [denominator excludes previous break-ups and other ineligibles]
Dates During Which Data was Gathered	Core Adult Profile: Earliest: 6/30/2003 Median: 11/2/2007 Latest: 3/20/2009 Public Affairs Profile: Earliest: 5/21/2005 Median: 9/25/2007 Latest: 10/19/2008	Earliest:2/21/2009 Median:2/23/2009 Latest: 4/3/2009	Earliest: 3/29/2009 Median: 4/23/2009 Latest: 3/4/2010	Earliest: 3/12/2010 Median: 3/18/2010 Latest: 6/8/2010	Earliest: 3/7/2010 Median: 7/2/2011 Latest: 9/12/2011	Earliest: 3/22/2011 Median: 4/8/2011 Latest: 8/29/2011	Earliest: 9/15/2011 Median: 5/9/2013 Latest: 11/6/2013	Earliest: 3/2013 Median: 3/2013 Latest: 11/2013
Order of variables in the Public Dataset	1	2	3	4	5	6	In Wave 4 supplement	In Wave 4 supplement
Associated Variable Prefixes	pp and pap	s# and q# for variables from HCMST main survey; plus many other variables without prefixes derived from these.	pp2_	w2_	pp3_	w3_	pp4_	w4_
Range of Variables, in Order in the dataset	ppage to pphhcomp11_member15_relationship	irb_consent to coresident Also note that grouped among the HCMST main survey variables are some variables relating to respondent’s race, religion, or sexual orientation that are based on the background demographic surveys, but are grouped here to be proximate to the same variables for respondent’s partner.	pp2_afterp1 to pp2_ppcmdate_yrmo	w2_deceased to w2_days_elapsed	pp3_pphhhead to pp3_newer	w2w3_combo_breakup to w3_nonmbtiming_month	pp4_ppeduc- ppa2009_how_often_services	w4_xqualified- w234_combo_breakup
Some Key Variables	* ppppcmdate_yrmo (year and month member completed Core Adult Profile) * pppadate_yrmo (year and month member completed Public Affairs Profile) Note that the background survey data applies only to the Respondent, not to Respondent’s Partner. The HCMST main survey supplies variables for the demographic background of Respondent’s Partner. * ppage (age incremented to yield age at the time respondent completed the HCMST main survey) * ppgender (respondent’s gender) * Information about up to 14 household members in addition to the respondent, numbered 2-15. pphhcomp11_member#_age pphhcomp11_member#_gender pphhcomp11_member#_relationship * children_in_hh (number of minors in respondent’s household) * ppmarit Marital status * ppincimp, hhinc Household income (the latter is a recode of the former). * ppwork (employment status) * ppeducat, ppeduc (educational attainment); see also respondent_yrsed, * respondent_race based on ppq14arace, pphispan, pprace_*  (race), ppethm. * papreligion, papevangelical (religion) * pppartyid3 (political party affiliation)  *ppnet (whether respondent had their own Internet access at home at the time of the pp survey. Note: most of the “no” households had Internet access provided by KN)	* HCM_main_interview_yrmo The year and month respondent finished the HCMST main survey * qflag identifies the 3,009 partnered respondents. The remaining 993 respondents did not have a spouse or partner at the time of the main survey. * recsource identifies the oversampled GLB respondents * papglb_status Identifies every respondent’s prior GLB status * weight1 Main weight for all respondents * weight2 Main weight for partnered respondents * respondent_race, partner_race * ppage (respondent age); q9 (partner age) * HCMST identifies not only respondent and partners’ current religions, but also the religion each was raised in, see: respondent_religion_at_16, respondent_relig_16_cat, partner_religion_at_16, partner_relig_16_cat which are derived from q8a, q8b, q8b_other; q13a, q13b, q13b_other * respondent_yrsed, partner_yrsed Educational attainment of respondent and educational attainment of current partner * respondent_mom_yrsed partner_mom_yrsed respondent’s mother and partner’s mother’s educational attainment *How, when, and through whom respondent and current partner first met, including: * q24_met_online, q32, q32_internet, how_met_online, either_internet, either_internet_adjusted (several measures of whether and how respondent met partner Online). * q24_* represent a variety of variables derived from the open-text question q24, “How Did You Meet your Spouse/Partner?” See online documentation “Notes on Coding of Open Text Q24” * The how did you meet variables indicate whether the person responsible for bringing respondent and partner together (for instance, a “friend”) was respondent’s friend, partner’s friend, or the friend of some other intermediary. See q24_R_friend,  q24_P_friend,  q24_btwn_I_friend * New to version 3.0 of the public data, q24_fam_* variables specify which family members played the intermediary role *Timing of first meeting, beginning of relationship, first cohabitation, first marriage, and/or first domestic partnership with current partner. See q21* as well as: * how_long_ago_first_met is years elapsed since respondent first met partner * how_long_relationship is current relationship duration * q17c, q17d, gender_attraction  (Sexual preference) * same_sex_couple (whether couple is a same-sex couple) * for Partner’s gender, see q4, q4_other, q5 * q12 (Partner’s political party affiliation), see also pppartyid3 (Respondent’s political party affiliation) * q15a1 and q15a2 (where Respondent was raised) * distancemoved_10mi Geographic mobility of respondent * parental_approval Parental approval or disapproval of respondent’s relationship with partner * relationship_quality Respondent’s assessment of relationship quality with partner	* pp2_afterp1 identifies the 3,693 cases who took the KN Core Adult Profile a second time. * pp2_ppcmdate_yrmo year and month of second Core Adult Profile plus all the relevant variables from Core Adult Survey for which new data would be interesting, such as: pp2_ppmarit, pp2_ppincimp, pp2_ppwork, pp2_ppnet and (new to version 3.0 of the public data) some background variables that you might not expect to change, but where the changes might be interesting, such as: pp2_ppethm (race) pp2_ppeduc (Note: unless otherwise noted, all Stanford-produced supplementary race, and education variables are derived from the original pp* variables because the first Core Adult Survey had the fewest missing values)	* w2_assigned (indicates which of the 3,009 partnered respondents from Wave I were attempted for the first follow-up survey) * w2_f1complete (indicates the 2,520 respondents who completed the first follow-up survey) *  w2_deceased, w2_multiname, w2_donotcontact (indicate various reasons why partnered respondents from the main survey might not have been assigned for the follow-up survey) *  w2_broke_up (indicates whether respondent is still partnered with their partner from the main survey, or not) * w2_x* are questions derived from earlier waves that were used to route respondents to the proper branch of the Wave 2 questionnaire; see Wave 2 survey instrument. * The actual questions that respondents saw are w2_q1- w2_q10	* pp3_newer identifies the 2,644 cases who took the KN Core Adult Profile a third time. * pp3_ppcmdate_yrmo year and month of third Core Adult Profile plus all the relevant variables from Core Adult Survey for which new data would be interesting, such as: pp3_ppmarit, pp3_ppincimp, pp3_ppwork, pp3_ppnet and  (new to version 3.0 of the public data) some background variables that you might not expect to change, but where the changes might be interesting, such as: pp3_ppethm (race) pp3_ppeduc (Note: unless otherwise noted, all Stanford-produced supplementary race and education variables are derived from the original pp* variables because the first Core Adult Survey had the fewest missing values) * interstate_mover_pp1_pp2, interstate_mover_pp2_pp3, and interstate_mover_pp1_pp3 record whether the subject lived in a different state at the time of the different core adult surveys.	*w3_xqualified identifies the 2,689 subjects who were qualified to take Wave 3, on account of not having previously broken up with partner, and partner not previously reported as deceased. w3_xpartnered, w3_xdeceased, and w3_multiname explain which cases were qualified for Wave 3. * w3_complete indicates the 1,960 cases who completed Wave 3. * w3_broke_up indicates which cases newly reported breakup in Wave 3. w2w3_combo_breakup indicates which cases were broken up or lost to follow-up across waves 2 and 3. * w3_x* are questions derived from earlier waves that were used to route respondents to the proper branch of the Wave 2 questionnaire; see Wave 2 survey instrument. * The actual questions that respondents saw are w3_q1- w3_q10, questions which are the same as the Wave 2 questions. * Wave 3 asked additional questions (not asked in Wave 2) of all newly reported break-ups, to determine year and month of the breakup, see: w3_mbtiming_year w3_mbtiming_month for married couples, and w3_nonmbtiming_year w3_nonmbtiming_month for unmarried couples.	*pp4_newer identifies the 2,035 subjects whose pp4 variables represent new information. Note that even if pp4_ values are not newer, the pp4 variable hold the earlier pp values (i.e., repeats of pp3 or pp2 values if pp4 is not newer information). In addition to the pp4 variables that repeat the pp3, pp2, and pp variables, we added a new retrospective 2009 religiosity variable, ppa2009_how_often_services, with its own date collected variable, ppa2009_services_yrmo	* w4_xqualified identifies the subjects who were qualified for Wave 4. * w4_x* are the variables that were used to determine what Wave 4 questions the subjects would see. In addition to the w4 variables that repeat w3 and w2 questions, there are some new questions: * w4_attractive and * w4_attractive_partner include the subject’s rating of their own and their partner’s attractiveness. Also, we have a Wave 4 repeat of the Wave 1 relationship quality question, * w4_quality
Questionnaire Codebook posted online	See Questions from Knowledge Networks’ prior demographic surveys	See Main Survey Instrument Wave I	See Questions from Knowledge Networks’ prior demographic surveys	See Survey Instrument Wave 2.	See Questions from Knowledge Networks’ prior demographic surveys	See Survey Instrument Wave 3.	See Questions from Knowledge Networks’ prior demographic surveys	See Survey Instrument Wave 4.