Page 122 - IJPS-1-1

Using panel data to examine pregnancy attitudes over time

fewer differences were statistically significant. Change in employment status was excluded
from the model estimating the likelihood of experiencing changes in trying to get pregnant,
and race/ethnicity from the pregnancy avoidance model. None of the interaction effects
between education and the other covariates were significant at 10 percent level, and were
thus not included. See Supplementary Online Materials A Tables A1 and A2 for full results
of the models. Fitted probabilities, which were calculated based on the models, are shown
below in Figures 1 and 2.
Figure 1 shows the fitted probabilities for changes in trying to get pregnant based on the
multinomial model. Again, transitioning into a stronger union or not changing one’s union
status was associated with starting rather than stopping trying; women in their late 20s
were more likely to start than stop trying; and mothers of infants and toddlers were rela-
tively likely to start trying to get pregnant. Those who had a college degree more often
started than stopped trying. Being childless or having at least three children was associated
with a higher probability to start trying. Hispanic and White women were less likely to
stop trying and more likely to start than other racial or ethnic groups.
Figure 2 shows the fitted probabilities of experiencing a change in pregnancy avoidance
based on the multinomial model. The directions of associations were similar to the model
where transitions in trying to get pregnant were studied for most variables. Moving to the
“next stage” in one’s union was associated with a relatively high (14 percent) probability
in transitioning into weaker pregnancy avoidance. Women in their late 20s and early 30s
had a higher probability of transitioning into weaker avoidance than other women. Women
whose youngest child was an infant had a markedly higher probability of transitioning into
weaker pregnancy avoidance compared to women with older children.
Unlike in the model measuring changes in trying, women who had less than a high
school education had relatively high probability of transitioning into a weaker avoidance
(14 percent), but women with a college degree were also more likely to transition into a
weaker than into a stronger avoidance (9 percent vs 5 percent). Women with at least two
children had a higher probability of transitioning into a weaker than stronger avoidance.
Constantly working full-time women were more often associated with transitioning into a
weaker avoidance than into a stronger one (Figure 2).
4. Discussion

While strong pregnancy avoidance and not trying to get pregnant was the norm for women
in our sample, our results show that pregnancy attitudes change for a non-negotiable mi-
nority of women over a relatively short period of time. Perhaps not surprisingly, pregnancy
avoidance showed more movement than efforts to get pregnant. Pregnancy avoidance has
a behavioral element; as many women who have a strong desire to avoid pregnancy are
likely to engage in practices to prevent this from happening, but it is less exclusive than
those women who were reportedly actively trying to get pregnant.
Women in the lowest level of education were relatively likely to transition to weaker
pregnancy avoidance, but less often into trying to get pregnant, whereas women with at
least college degree had both higher likelihood of transitioning into weaker avoidance and
starting to try. This may reflect different strategies of planning childbearing. According to
TCA, after having decided to start a family, women with higher education are more likely
to change their behavior beyond just not using contraception, for instance, by optimizing
the timing of intercourse. Women from less advantaged backgrounds may take a more in-
formal approach to childbearing, for example, accepting a pregnancy even when it is un-
planned or alternatively, stopping contraception to show commitment to their partner
(Johnson-Hanks, Bachrach, Morgan et al, 2011). Thus, there may be a higher likelihood of

International Journal of Population Studies | 2015, Volume 1, Issue 1 116