Experimental Design
I. DATA COLLECTION
This survey is programmed in Qualtrics and distributed via Prolific. Subjects are pre-screened to only include (i) parents, (ii) from the US, (iii) who are fluent in English and have English as their primary language, and (iv) have an approval rating of at least 85 %. Further, it was specified that half of the subjects should be male and half female. The survey consists of two parts.
Part 1
In part 1, a large number of questions are asked in the context of the shaping of their child’s attitudes and character traits. For comparability reasons, parents are asked to respond to each question thinking about their child. For each subject, four attitudes and character traits out of a total of seven are randomly selected (Patience, Altruism, Risk Attitude, Work Ethic, Political Attitude, Religiosity, Universalism). For each of the four randomly selected attitudes/character traits, subjects are then asked:
DeFacto:
How much do (i) parents, (ii) other relatives, (iii) schools/teachers, (iv) peer group/friends and (v) social media/influencer shape these traits in their first child?
Time:
For each of the five agents or groups of agents above (i – v) that parents believe to shape a trait at least somewhat: At what age of their first child do parents think the shaping takes place predominantly?
Preference:
For each attitude/character trait that parents think is at least somewhat shaped by them: How strongly do parents themselves want to shape each trait?
OptimalLevel:
For each attitude/character trait: What is parents’ belief about their confidence in knowing what is the ‘optimal’ level their child should possess?
Goal:
For each trait that parents want to shape at least somewhat: With which goal in mind do parents themselves want to shape the trait?
Method:
For each trait that parents want to shape at least somewhat: By which means do parents want to shape the trait?
Last, in case Risk Attitude was one of the randomly selected attitudes/character traits, parents are asked: “When you were answering these questions, what kinds of risks were you thinking about? Please write down a few key words.” (free-text, 60 characters maximum).
Part 2
In part 2, data to test two theoretical assumptions – “Imperfect Empathy” and “Substitutability” – are collected.
Part 2.1: Imperfect Empathy
Subjects are given two stylized job descriptions (in short, job 1: stable, predictable, more family-friendly; job 2: fast-paced, higher salary, more career-oriented), and are asked:
ImpEmpParent:
Looking back now, which job do you wish you had chosen as your first job?
Then, subjects are told to assume their first child would want to choose the other job as their first job, and they are asked:
ImpEmpChoice:
Which job would you choose for your first child as their first job?
ImpEmpReason:
Why would you choose this job for your first child?
ImpEmpLifeCycle:
Maybe people change their preference regarding their first job over the course of their life. How relevant do you think that such a change is in the context of your previous answers about your and your first child’s job preferences?
ImpEmpWorldChange:
Maybe the world has changed between when you chose your first job and when your first child is choosing their first job. How relevant do you think that such a change is in the context of your previous answers about your and your first child’s job preferences?
Part 2.2: Substitutability
Between-subjects, treatment variation is introduced: In Liberal, subjects are told to imagine that their first child’s school and friends emphasize liberal values. Likewise, in Conservative, subjects are asked to imagine that their first child’s school and friends emphasize conservative values.
Then, subjects are asked:
SubstChoice:
Assuming this scenario to be true, which values would you emphasize in the interaction with your child?
SubstIntensity:
How much effort would you put in emphasizing these values?
Demographic and Socio-Economic Covariates
At the end, these demographic and socio-economic covariates are elicited: birth year of subject; gender of subject; residence (US state) of subject; employment status of subject; whether subject was born in the US; ethnicity of subject; education level of subject; whether subject is still in a relationship/marriage with the other parent; if not, whether they are in a relationship/marriage with a new partner; (if applicable) whether other parent or new partner has had more influence in shaping the subject’s first child’s attitudes and character traits; multiple-choice personality description of subject, the other parent and (if applicable) the new partner; how good the relationship between parent and first child is in general; parenting style subject used with first child; parenting style the other parent used with first child; (if applicable) parenting style the new partner used with first child; general division of parenting responsibilities between subject and the more relevant partner (according to question above); division of parenting responsibilities specifically in shaping the subject’s first child’s attitudes and character traits between subject and the more relevant partner (according to question above); number of siblings of subject; parenting style father of subject used; parenting style mother of subject used; type of school first child attended; how liberal vs. conservative subject wants the first child to be (used as PolTargetChild in analysis of part 2.2 , see below); political orientation of school/teachers of first child; political orientation of friends/peer group of first child; first child birth year; first child gender; multiple-choice personality description of first child; gender of subject’s other children; perceived importance of life areas (e.g., family, leisure, work etc.) for subject; annual household income; agglomeration type; political attitude (liberal vs. conservative) of subject, other parent and (if applicable) new partner.
II. TREATMENT CONDITIONS
There are two treatment conditions. These conditions only concern Part 2.2 “Substitutability.”
Liberal: Subjects are asked to imagine their first child’s school and friends emphasize liberal values.
Conservative: Subjects are asked to imagine their first child’s school and friends emphasize conservative values.
Assignment to the Liberal and Conservative conditions was specified to be approximately balanced in Qualtrics, with half of the subjects allocated to each condition; however, exact numbers may differ slightly due to randomization and exclusions.