Page 134 - IJPS-11-3
P. 134

International Journal of
            Population Studies                                                    Drivers of reproductive delay in the UK



            people trade off more of one attribute for less of another   matrix of choice sets large enough to force participants to
            (Cleland et al., 2018; Louviere et al., 2010). An attribute   make  trade-offs  between  the  attributes  while  also  being
            is a component that contributes to making a decision: it   few enough to not be overly cognitively burdensome or
            is a variable such as financial security, partnership status,   time-consuming. The D-efficient design also minimizes
            and employment status. Each attribute has levels which   the standard errors of the parameter estimates produced
            can be either qualitative, such as partnership status with   from a limited number of choice sets (Rose & Bliemer,
            levels that might be “single,” “married,” “cohabiting,” or   2009). In this study, the DCE was made up of 16 choice
            quantitative, such as salary or price, but in this case, it is the   sets. The nature of the design and the presentation of the
            timing of the baby’s arrival in months. This continuous-  DCE  instrument  means  that  people  make  choices  quite
            level attribute can be divided into discrete components   quickly and more intuitively than they do when answering
            for presentation in the DCE instrument and then used to   direct questions (Hensher, 2014). By doing so, they reveal
            calculate an estimate of willingness to pay for each attribute.   their actual preferences rather than overthinking the
            The underlying assumption is that individuals make   answers or being otherwise influenced, as may be the case
            choices based on the attributes that make up the “product,”   when completing regular surveys. Statistical analysis of
            in a way that maximizes utility, and by making trade-offs   the resulting choices reveals the underlying preferences of
            between the attributes (Kjaer et al., 2006). In the application   individuals, or their “utility function;” their value beyond
            of a DCE, individuals are asked to make a series of choices   their cost.
            between alternative hypothetical scenarios (called choice
            sets); with each choice set being made up of a number of   2.2. Willingness-to-pay
            attributes that are split into levels. Table 1 shows an example   DCEs typically model a price or salary variable to measure
            of a choice set as it looked when presented in the online   the maximum amount a consumer is willing to pay (WTP)
            DCE for this study. Before choosing from each choice set,   for a product feature or how much a potential worker
            participants are given qualitative information about what   values characteristics of a job as measured in salary, that is,
            each attribute means. It is impossible to ask participants   how much they would need to be paid to take up the job.
            to make choices from a fully-factorial design where all   In this study, WTP was modeled as the maximum amount
            possible  choices  are  presented as  this  would  mean,  for   of time, in months, an individual is willing to forgo to have
            example, for a DCE with five attributes of three levels each,
            245 possible choices would be generated (i.e., 3 ). Instead,   their needs met before having a baby. This can also be
                                                  5
            a D-efficient experimental design is applied, producing a   interpreted as the time that can potentially be recuperated
                                                               if the barriers to reproduction were lifted.
            Table 1. An example of a choice set, presented as part of a   2.3. Data
            battery of 16, to university-educated women in the UK study
                                                               The successful design of the DCE depends on the quality
                              Scenario A       Scenario B      and relevance of available data on the experiment attributes,
            Partner support  My partner is fully involved  My partner is   in this case, the barriers to reproduction. In the UK, there
                         and happy to share the   supportive but not   are no recent data on this and so we conducted a series of
                         childcare 50/50.   very hands-on with   focus groups to learn what people in the UK currently feel
                                            childcare.         is preventing them from attaining their fertility goals. The
            Career costs  I’ve just started a new job   My career is well-  focus groups generated a rich corpus of qualitative data
                         which is promising but will  established and I am
                         take some time to establish   supported at work.  from men and women, aged 23 – 43, university-educated
                         myself.                               and not, producing the most current and nuanced data on
            Friends      My closest friends are all   My friends are not   what the barriers in the UK currently are. The qualitative
                         planning babies right now.  very baby-oriented   leg of this study is described fully elsewhere (Brough &
                                            and many don’t want   Sheppard, 2022). A thematic analysis was applied to the
                                            kids.              focus group data and the results were used to construct a
            Finances     The household finances are   The household   shortlist of four primary barriers (i.e., the DCE attributes)
                         secure and I have enough   finances are a bit   for each demographic group (men, women, university, or
                         savings.           stretched and there is
                                            no savings buffer.  not) with which to design the experimental instruments.
            Baby timing  The baby will be born 18   The baby will be born   Each group expressed different priorities and so there was
                         months from now.   3 years from now.  a different DCE for each group.
            Which scenario is                                    The DCEs were administered by SurveyEngine (a
            better?                                            multinational company), a company that specializes in



            Volume 11 Issue 3 (2025)                       128                        https://doi.org/10.36922/ijps.3600
   129   130   131   132   133   134   135   136   137   138   139