In many democracies, the debate over restrictions on abortion has been
the subject of heated controversy. Across the globe, from Switzerland to
South Africa to Australia to Taiwan, legislatures have in recent years
struggled with the issue, with the result being more liberal legislation
in most cases (Tribe 1992). The Center for
Reproductive Rights, which tracks abortion legislation worldwide, notes
that only five countries have adopted more restrictive abortion laws since
1995—the United States, the Russian Federation, Hungary, Poland, and
El Salvador. These countries have limited women's reproductive
rights, in some cases significantly. What explains these moves to restrict
women's access to abortion?
For two of these countries, the United States and Poland, there is
comparable data on public attitudes on abortion since the early 1990s. The
comparison of the two countries is useful, for both have experienced a
heated debate on abortion in which religious conservatives have marshaled
arguments about the sanctity of life as well as sexual morality. In both
countries, national constitutional courts have intervened in legislative
efforts, though in different directions. In the United States, lawmakers
repeatedly passed legislation banning intact dilation and extraction
(so-called partial birth abortion) during the 1990s, but President Bill
Clinton vetoed it. In 2003, Congress again passed the ban and President
George W. Bush signed it. Federal appeals courts have held the law
unconstitutional, but the Supreme Court has yet to rule. In Poland, the
Parliament passed perhaps the most restrictive abortion measure in Europe
in the immediate aftermath of the fall of the communist regime. In 1996,
the Parliament passed a law liberalizing abortion, but the Constitutional
Tribunal overturned it. In 1997, a new Parliament further restricted
abortion access.
In both countries, the political debate over abortion has been
animated, and the major political parties have staked out conflicting
political positions (Adams 1997; Girard and Norwicka 2002). In the United States,
abortion has become an important voting cue, and has motivated
unconventional and occasionally violent acts on the part of pro-life
activists (Abramowitz 1995; Cook, Jelen, and
Wilcox 1994a, 1994b;
Maxwell 2002). In Poland, abortion debates have
involved religious groups, physicians, and women's groups (Girard and Nowicka 2002). Several Polish parties of
the Left, coalescing in the Democratic Left Alliance in 1993, have
aggressively pushed for liberalization of an extremely restrictive law
passed with the support of the Catholic Church immediately following the
fall of the communist regime in 1989 (Byrnes
2002). The church is a powerful political actor in Poland, and its
legitimacy was enhanced by its role in maintaining national identity
during the period of communist domination in the post–World War II
era (Davies 1982). However, attempts by the
church to protect its special position in Polish society, and,
specifically, to restrict abortion rights, have caused many Polish
Catholics to regard the church as “too powerful” (Eberts 1998).
In both countries, the abortion issue has been framed in recent years
around religious considerations, around fetal rights, and about other
issues, but frames regarding the impact of abortion and childbirth on
women have been largely absent. Indeed, given the crucial nature of the
abortion issue to women, the debate in each nation has been remarkably
silent on the relationship between reproductive rights and women's
bodily integrity, women's personal autonomy, and women's
citizenship rights. In the United States, the debate has centered on
competing rights to choose abortion or rights to fetal life, but there has
been relatively little discussion of the gendered consequences of abortion
policies on women's lives. The abortion debate in the United States
has largely focused on the exercise of individual rights, but has not
emphasized the gendered social contexts in which issues of reproductive
freedom necessarily arise (Saletan 2003).
Similarly, in Poland, public debate on the abortion issue has emphasized
the possibility of fetal life, and has specifically centered on the Roman
Catholic concept of natural law (Dillon 1996).
As in the United States, and despite the efforts of some women's
groups, public discourse on abortion in Poland is highly abstract and
somewhat divorced from the gendered context in which decisions about
reproductive freedom are often made (Girard and Nowicka
2002). In other words, the framing of the debate around changes in
abortion law in each country has been silent on women and women's
lives, and has been effectively degendered.
These two countries are outliers, then, in an international trend
toward liberalization of abortion laws. Have legislatures enacted
restrictive laws because public opinion in these countries has become less
supportive of abortion rights? Or have religious groups lobbied
successfully to enact their preferences in the absence of public demand?
The degendering of the abortion issue has shifted the grounds of the
public debate. Has it had the same effect on actual public opinion?
Public opinion on abortion can be both a consequence and a cause of
policy change. In the United States, key court decisions that have
legalized and later limited abortion rights have stimulated the
mobilization of public opinion (Franklin and Kosaki
1989; Freedman 2004; Wlezien and Goggin 1993). Yet changing levels of
public support for legal abortion clearly alter the willingness of elected
officials to limit access to abortion (Norrander and
Wilcox 1999). Indeed, Wetstein (1996) has
shown that in the United States, public opinion is generally the strongest
predictor of access to abortion services, despite the fact that popularly
elected branches of government operate under substantial constraints that
are judicially imposed.
SOURCES OF OPINION CHANGE
Aggregate public support for legal abortion can change for three
distinct reasons. First, the composition of the population may change over
time. For example, it is well known (Cook, Jelen, and
Wilcox 1992; Luker 1985) that among
women, participation in the paid labor force is associated with more
permissive attitudes toward abortion. Aggregate abortion attitudes may
move in a pro-choice direction, even if the relationship between female
labor force participation and abortion attitudes remains constant, if the
proportion of employed women rises over time. Similarly, an increase in
the overall religiosity of a population could occasion an increase in
pro-life attitudes, even if the relationship between abortion attitudes
and religiosity remains unchanged. Hence, compositional changes can effect
attitudinal changes.
Second, the sources of abortion attitudes may change as different
groups of citizens change their opinions on abortion. Political and
religious groups may become mobilized or demobilized as the issue evolves.
For example, the dynamics of public opinion in the United States have
changed as some groups of Roman Catholics have become more pro-choice and
evangelical Protestants have moved in a more pro-life direction (Sullins 1999). In Poland, although the proportion of
Roman Catholics has remained virtually constant, the political role of the
hierarchy has changed substantially, in response to rapid alterations in
political institutions and political culture (Byrnes
2002). Indeed, some observers have suggested that the prestige of
the church has declined during the transition from communism to multiparty
democracy, which could reduce the magnitude of the relationship between
Catholicism and attitudes toward abortion (Byrnes
2002; Jelen and Wilcox 1997; Taras 2003). Structural changes, therefore, can cause
aggregate attitudinal changes.
Finally, abortion attitudes can change because of events or arguments
that affect all elements of society roughly equally. The political debate
can frame the abortion issue in ways that change the opinions expressed in
surveys (Freedman 2004; Wilcox and Riches 2002). Substantial ambivalence
characterizes public attitudes on abortion; many citizens value both fetal
life and women's autonomy (Alvarez and Brehm
1995; Cook, Jelen, and Wilcox 1992). For
many, this ambivalence is expressed by uncertainty over whether to permit
abortion under some circumstances, and the precise circumstances that are
troubling vary across individuals and social groups as well. When prompted
to offer an opinion on abortion in these difficult cases, citizens access
considerations that include the frame of the recent policy debate. As
issue framing shifts over time, attitudes in the mass public are likely to
shift in response.
Social movements often compete to frame policy debates in ways that
help to mobilize their own members and to persuade ambivalent citizens
(Franceschet 2004; Snow and
Benford 1992). In the United States, pro-choice groups succeeded in
framing the abortion debate over the question of “who decides”
when it seemed possible that Roe v. Wade might soon be repealed,
thereby allowing states to restrict abortion rights significantly (Wlezien and Goggin 1993). By the middle of the 1990s,
however, conservatives had succeeded in reframing the issue around graphic
descriptions of one particular abortion procedure, and the media adopted
the pro-life group's label of “partial birth” abortion to
describe the procedure. In 1995, the issue became more salient: A
Lexis/Nexis search of major newspapers turned up 98 mentions of the
term “partial birth abortion” in major newspapers. Two years
later the number of citations topped 500; the figure remained high
throughout 2000. It is interesting that this frame became less visible in
subsequent years, with only 28 mentions in 2002.
In Poland, debate over abortion has been more general in tone, with
the Catholic Church asserting that abortion violates the sanctity of human
life. Although there has been a heated debate about the role of the church
in politics, and the prestige of the church has declined somewhat, several
factors led the church to have special influence on this issue. First,
abortion was widely practiced during Soviet occupation, and the church is
widely credited with helping end that period of Polish history. Second,
Polish Pope John Paul II was very popular during this period, and he
articulated a vision wherein Poland had a unique role as the guardian of
Catholic values in the new Europe (Byrnes 2002).
Moreover, although scholars report that the prestige of the church has
declined since the fall of communism, the Polish General Social Survey
data show that confidence in the church was higher at the end of the 1990s
than at the start, and that confidence was far higher than in the
government or in other national institutions.
Of course, all three of these effects—compositional, structural,
and period—can occur simultaneously, and their mix may well vary
from country to country and across time. In this article, we seek to
estimate and compare the direction and magnitude of compositional,
structural, and period effects on abortion attitudes for Poland and the
United States. Although we are not able to analyze shifts in social
movement framing of the abortion issue, or changes in the involvement and
public activism of relevant interest groups and movement actors, we
examine shifts in public opinion concerning abortion, and the potential
reasons for them, as a starting point for understanding why nations might
become more restrictive in abortion policies over time.
DATA AND METHOD
Data for this study were taken from the General Social Surveys in the
United States from 1991 to 2002, and from the Polish GSS from 1992 to
2002. Respondents in both countries were asked an identical battery of
questions about whether abortion should be legal if the woman's
health was in danger, if there was a chance for a serious defect in the
baby, if the woman became pregnant as the result of rape, if the family
was poor, if the woman was single and did not wish to marry, or if the
family did not want more children. The abortion scale is a simple count of
the number of affirmative responses, ranging from 0 (support for abortion
under none of these circumstances) to 6 (support for abortion under each
of these situations). This measure has been shown to be quite reliable in
both Poland and the United States (Jelen and Wilcox, 1997).
TRENDS IN ABORTION ATTITUDES
The data in Figure 1 show an impressive
level of stability in abortion attitudes overall (Converse and Markus 1979; Sharpe
1999; Wilcox and Norrander 2002). In each
year, in each country, the mean score on this six-point scale rounds to 4.
Yet there are also parallel trends, with a general decline in support for
legal abortion, followed by a small increase in support in 2002.
Poland/U.S. mean abortion attitudes.
For most years, Polish respondents are slightly more supportive of
legal abortion than their American counterparts. Support for legal
abortion in Poland was highest in 1993 (the year in which the Polish
legislature passed a highly restrictive abortion law) and appears to have
declined monotonically until an abrupt rise in pro-choice attitudes was
observed in the 2002 Polish General Social Survey data.
In the United States, the decline in pro-choice attitudes appears to
have begun in 1996, although we have shown elsewhere that the relatively
high level of support for legal abortion represents an increase over the
period immediately preceding Webster (Wilcox
and Jelen 2002). The decline in support for legal abortion in Figure 1 is statistically significant, but small,
dropping from a high of 4.0 in 1994 to a low of slightly less than 3.7 in
2000.
These small changes in attitudes as measured in the GSS translate,
however, into much larger changes politically. In the United States, the
debate over abortion has been symbolically rendered into a clash between
“pro-choice” and “pro-life” camps. Relatively
small changes in the GSS scale translate into much larger changes in
self-identities as pro-life or pro-choice: Gallup polling that asks
respondents whether they are more pro-life or pro-choice shows a decline
in pro-choice identities from 56% to 46% between 1995 and 2002, and an
increase in pro-life identities from 33% to 46%.
The similar distributions of overall attitudes in these two countries
are mirrored by a similarity in the overall structure of attitudes. In
both countries, support for legal abortion is much higher for physically
traumatic circumstances such as dangers to the woman's health, fetal
defect, and rape, and lower for socially traumatic circumstances such as
poverty, single motherhood, and a family that wants no more children.
Although we will see different predictors of abortion attitudes in each
country, these data suggest that the issue is perceived in a similar way
in both countries.
Compositional Changes
As we noted, the changes in abortion attitudes in Figure 1 might be attributed to demographic and
attitudinal changes in the populations of Poland and the United States.
For example, increases in religiosity or religious observance, or in the
average age of Polish or American mass publics, might increase pro-life
sentiment. The relationship between, for example, religiosity and abortion
might be quite stable for the period under consideration, but support for
legal abortion might nonetheless decline due to increases in personal
religious intensity. Conversely, it is possible that changes in the
aggregate demographic or attitudinal characteristics of the population of
either country might lead one to expect increases in the frequency of
pro-choice attitudes (Wilcox and Norrander
2002). Increases in education, female participation in the paid
labor force, or greater acceptance of sexual activity outside of marriage
might create increased support for reproductive freedom. If this is the
case, the small changes observed in Figure 1 may
represent more profound pro-life forces in public opinion in the two
countries, which are partially concealed by compositional changes in
characteristics associated with pro-choice attitudes.
We estimate the effects of compositional change for both countries in
Figures 2 and 3. In
each figure, the observed mean on the abortion scale is compared with a
simulated mean. The simulated mean represents an attempt to isolate the
effects of composition change in the population, by holding constant
changes in the political environment, or in the relationships between
abortion attitudes and variables that have been shown to be strong
predictors of such attitudes.
Comparison of observed and simulated means on abortion scale:
Poland.
Comparison of observed and simulated means on abortion scale: United
States.
The simulated means are generated by a fairly straightforward
procedure. For each year, in each country, we estimate a multivariate
ordinary least squares equation for each country, in which the dependent
variable is the six-point abortion scale. The variables in the equations
represent a mixture of attitudinal and demographic items. The Polish
equations include attitudinal items measuring attitudes on euthanasia,
sexual morality, public feminism, and subjective religiosity. The Polish
equations also include items that measure church attendance, marital
status, number of children living at home, whether the respondent worked
full time, self-identification as a homemaker, age, sex, and respondent
educational attainment. The equations estimating abortion attitudes in the
United States are identical, except that the U.S. equations also include
measures of Roman Catholicism, membership in an evangelical Protestant
denomination, respondent race, and respondent attitudes toward the Bible.
(For details of the dependent and independent variable, see Appendix.) The
lack of religious and racial diversity in Poland made the inclusion of
these items impractical in the analysis of Polish attitudes. For both
countries, the explanatory power of the models is quite impressive, with
adjusted R2 ranging from .27 to .41.
The next step is to reestimate the OLS models for each year, taking
the slope associated with the first year in the series (1991 in the United
States, 1992 in Poland), multiplying the slope by the mean value for each
respective variable for each year, and adding or subtracting (depending on
the sign of the relationship) the resulting products from the intercept
(constant) for the first year in each series. In using this procedure, we
seek to examine a counterfactual question: What would the distribution of
abortion attitudes in Poland and the United States be if the only changes
allowed were changes in the marginal distributions of the independent
variables? (See Chandler et al. 1994 for an
illustration of this procedure.)
In Figure 2, we compare the observed and
simulated means for Poland. As these data show, the effects of
compositional change in Poland, if isolated, exert strong pro-choice
forces. Although our primary interest is in the general pattern of
changes, we are able to offer tentative assessments of the effects of
particular variables. Inspection of the OLS models on which this figure is
based suggests that the main changes are greater acceptance of sexual
activity outside of marriage, increases in female participation in the
paid labor force, and slight declines in church attendance. In each case,
this is a secular trend that is relatively consistent over the entire
period of the study, and analysis of variance suggests that the linearity
of the trend is statistically significant at .01.
For each year, aggregate abortion attitudes predicted by compositional
change in the population are much more permissive than those actually
observed. The decline in predicted pro-choice attitudes in 1995 appears
attributable to a temporary increase in church attendance. Conversely, the
increase in observed and predicted pro-choice attitudes in 2002 seems
based on an increase in gender role egalitarianism.
The pattern found in the United States seems to offer a stark
contrast. The simulated and observed means on the abortion scale appear to
track each other quite closely, and diverge only between 2000 and 2002.
The slight decline in predicted and observed change in the United States
between 1996 and 2000 appears to be attributable to rises in the
percentage of the U.S. population reporting affiliation with evangelical
denominations, and by slight increases in subjective religiosity. Again,
these changes are statistically significant.
Thus, compositional changes are pushing Polish attitudes in a more
pro-choice direction, but exerting no net effects on attitudes in the
United States. It seems unlikely, therefore, that demographic and
attitudinal changes can account for the observed pro-life trends in Poland
and the United States. Indeed, the results of Figures 2 and 3 suggest that
compositional changes among the Polish and American populations in the
last years of the twentieth century should have resulted in a slight
increase in pro-choice sentiment in the United States, and a much larger
increase in Poland.
Structural Changes
We next consider the possibility of structural changes in abortion
attitudes in the two countries. That is, we seek to estimate the predicted
change in abortion attitudes that would have taken place in Poland and the
United States if the distributions of demographic and attitudinal
variables in the respective populations had remained constant, but the
relationships between those characteristics and abortion attitudes had
been allowed to change. For example, a mobilization of evangelical
Christians (as is observed in the United States) or a possible
demobilization of Roman Catholics in Poland might occasion changes in
aggregate levels of support for legal abortion, even if the proportion of
U.S. evangelicals or Polish Catholics had remained constant.
In order to estimate the simulated structural changes, we repeat the
simulation procedure used to estimate predicted compositional changes.
However, in this exercise, we hold constant the mean values on all
independent variables in the OLS models to their value in the first year
of the series (1991 in the United States, 1992 in Poland) and multiply
each constant mean by its associated coefficient for each year. We then
add or subtract the resulting products from the constant associated with
the first year in the series. Again, we seek to test a counterfactual:
What would be the aggregate distribution of abortion attitudes in each
country if only the relationships between the abortion scale and the
predictor variables in the OLS models were allowed to vary, and if
composition and period effects were held constant?
The results of this exercise reveal stark contrasts between the two
countries, as can be seen in Figures 4 and 5. In Poland, the simulated structural effects on
abortion attitudes are strongly pro-life. Inspection of the OLS models on
which Figure 4 is based suggests that much of
the change in simulated Polish abortion attitudes is caused by a decline
in the relationship between the support for abortion and respondent
beliefs about the morality of nonmarital sex. More striking is the
increasing strength of the relationship between the euthanasia item and
the abortion scale. This finding suggests that Polish respondents are
increasingly likely to regard abortion as “a matter of life and
death” and to consider the possible humanity of the fetus in their
abortion attitudes, and less likely to consider abortion as a matter of
sexual morality. Conversely, the slight increase in simulated pro-choice
attitudes, and the slightly smaller observed change that began in 1997,
can be attributed to an increase in support for gender equality in the
paid labor force.
Comparison of observed and simulated slopes on abortion scale:
Poland.
Comparison of observed and simulated slopes on abortion scale: United
States.
By contrast, the pattern for the United States in Figure 5 suggests that the structural aspect of
attitude change on abortion has changed in a strongly pro-choice
direction. The relationship between subjective religiosity, church
attendance, and self-identification as a Roman Catholic with abortion
attitudes has declined. These structural pro-choice changes are offset to
some extent by striking increases in the relationship between the abortion
scale and attitudes toward the authority of the Bible, and with respondent
membership in evangelical Protestant denominations. Moreover, there have
been slight declines in the relationship between abortion attitudes and
respondent opinions about nonmarital sex, which suggests that the common
American commitment to personal liberty is a declining source of
pro-choice sentiment in the United States.
To test the statistical significance of these trends in structural
effects, we calculated a z score based on the unstandardized regression
coefficients in the first and last years for each country, and the
standard errors. Jacob Cohen and Patricia Cohen (1983) provide the statistical justification for the
procedure. Most of the structural changes do not rise to the level of
statistical significance. In Poland, however, the role of attitudes toward
euthanasia in explaining abortion attitudes rises significantly between
1991 and 2000, before declining slightly in 2002. In the United States,
the declining impact of subjective religiosity is also significant.
Thus, the pattern of group mobilization and demobilization is quite
different in the United States and Poland. In Poland, attitudes toward the
sanctity of human life—for which the marginal distributions are
extremely stable—have become a strong source of pro-life sentiment
in Poland. The net result is that changes in the structure of public
opinion in Poland have exerted substantial pressure in the pro-life
direction. In the United States, the pattern of group mobilization and
demobilization appears to have exerted strong pro-choice forces. Although
the period of 1991–2002 has witnessed the mobilization of doctrinal
and denominational evangelical Protestants in a pro-life direction, this
trend is more than offset by a declining relevance of Roman Catholicism,
as well as behavioral and attitudinal religiosity.
Overall, then, structural changes are pushing U.S. public opinion
toward more support for abortion rights and Polish opinion toward more
conservative opinions. In the case of the United States, there is no
counterbalancing compositional trend; in the case of Poland, compositional
effects are pushing the country toward more liberal opinion.
Period Effects
It is perhaps instructive to consider other sources of attitude change
beyond the characteristics of individual citizens. In Table 1, we seek to model period effects, which
represent changes in abortion attitudes that cannot be attributed to
changes in the marginal distributions of predictor variables, or to
changes in the relationships between abortion attitudes and such
variables.
Table 1: Multivariate models of support for legal abortion, by country
(OLS regression; entries are unstandardized regression
coefficients)
In Table 1, we present three multivariate
models of abortion attitudes. One model presents an estimated OLS equation
for Poland, while “United States Model 1” contains the results
of a comparable model for the United States. “United States Model
2” contains a more fully specified model, which includes respondent
race, respondent religious denomination, and respondent attitudes toward
the Bible (none of which vary meaningfully among the Polish sample).
For present purposes, we are most interested in the coefficients for
dummy variables associated with each year of the survey in question. We
take these to be estimates of the period effects associated with that
year, when compared to the first year in the series (1991 for the United
States, 1992 for Poland). As these data show, there are no statistically
significant period effects for either country until the late 1990s.
Beginning in 1997 (for Poland) and 1998 (for the United States), the
estimated period effects are significant and negative. Each successive
year has significantly more conservative attitudes in each country than
the baseline first year.
It would be inappropriate to overestimate the magnitude of these
period effects. Nevertheless, the effects are highly significant, even in
the face of very elaborate multivariate controls, and as noted, these
relatively small changes in attitudes as measured by the GSS produce
larger political changes. Thus, both countries are experiencing a downturn
in support for legal abortion. In the United States, these period effects
occur despite changes in structural relationships that are pushing opinion
in a more liberal direction. In short, the true magnitude of the period
effects is disguised by these other changes.
The observed changes are consistent with our earlier discussion of the
framing of the abortion issue. In all societies, the abortion issue poses
a trade-off between key values, but the nature of the political discourse
varies across societies and, therefore, helps to frame the debate (Feree et al. 2002; Kelley, Evans,
and Headey 1993). It seems likely that the period effects in the
United States captured in Table 1 are caused by
sophisticated reframing of the abortion debate by pro-life leaders and
interest groups. William Saletan (2003) has
suggested that pro-life leaders were able to co-opt the frame of personal
autonomy by arguing that abortion is ultimately a decision that affects
entire families. In answer to the question “who decides?”
interest groups and legislators responded that “families
decide,” and they used a powerful pro-choice frame to garner support
for such restrictions as parental consent for abortions performed on minor
children. Second, the 1990s witnessed an emphasis on “partial
birth” abortion, as numerous state legislatures passed legislation
prohibiting the procedure. The U.S. Congress passed such legislation in
2003. It may be the case that the rhetorical emphasis on partial birth
abortion among pro-life groups, which emerged in the 1990s, is the most
visible manifestation of a strategy among opponents of legal abortion, one
that emphasizes the most graphic aspects of abortion. Rather than
asserting that “life begins at conception,” pro-life groups
seem to have shifted their focus to the human status of late-term fetuses.
Finally, and most recently, the Republican Congress has passed, and
President George W. Bush has signed a bill, that would provide legal
protection for “unborn persons” who are injured or killed
during the commission of crimes against the woman. These rhetorical
strategies represent the emergence of plausible pro-life frames, which
appear to have had an affect on public opinion.
In Poland, abortion was legal and free during the communist years, and
contraception was both erratically available and fallible. As a
consequence, some 180,000 abortions were performed annually during the
final years of the regime. Among many Poles, frequent and easily available
abortion was among the most heinous abuses of the Stalinist era (Byrnes 2002). When Poland gained its independence,
Catholic groups lobbied hard both in the legislature and within the
medical profession to criminalize the procedure. In 1992, the medical
profession voted to bar abortions, and in 1993, the Parliament passed an
anti-abortion measure that made most abortions illegal, except in the case
of danger to the physical health of the woman, severe fetal defect, and
pregnancies that are the result of criminal acts. The law was liberalized
in 1994, only to be overturned by a presidential veto. Parliament passed
another liberalized law in 1996, which was held to be unconstitutional by
a decision by the Constitutional Tribunal (Girard and
Nowicka 2002). Indeed, the grounds for both the veto and the
judicial reversal appear to be a statement guaranteeing the “right
to life” in the Polish constitution (Byrnes
2002). Although activist groups argued the case for women's
interests in some popular media (see Girard and Nowicka
2002), women's rights do not appear to have been a
consideration in the Constitutional Tribunal's decision.
In the 2001 elections, the Democratic Left Alliance promised to relax
restrictions on abortion, but to date there has been no legislation. The
abortion issue remains on the agenda of the political Left in Poland, but
the 1993 law has remained in force. Poland's restrictive abortion
policy has made that country the object of much international criticism
and has affected the domestic debate over Poland's entry into the
European Union (Kaiser Daily Reproductive Health Report
2003).
It seems likely that this extensive, highly visible political activity
has contributed to the secularization and simultaneous degendering of the
abortion debate. In the years immediately following the decline of the
communist regime, abortion rhetoric was dominated by the Catholic Church,
which provided a highly theological, “natural law” rationale
for opposing legal abortion (Dillon 1996).
However, the interplay among political parties, the national legislature,
and the judiciary may have forced political actors to provide nonreligious
rationales for their positions. Further, the interjection of the abortion
issue into the question of Polish admission to the EU may have had the
effect of mobilizing nationalist sentiment around the visible symbol of
Polish identity. If this is the case, the fact that abortion has become a
contested issue in the secular world of politics may have legitimized the
public expression of pro-life attitudes, even among those who are not
swayed by the dominant frame of the church.
While a thorough analysis of political rhetoric surrounding the
abortion issue is well beyond the scope of this article, the possibility
that the public framing of the abortion issue in Poland has occasioned
movement in the pro-life direction poses a fascinating set of questions
for future research. It is worth noting that the legal limitations in
Poland are significant. The 1993 legislation prohibited abortions with the
following exceptions (Byrnes 2002):
- when the pregnancy endangers the woman's life, as certified by
a panel of physicians;
- when the pregnancy is the result of rape or incest, as certified by
a prosecutor;
- when prenatal tests show that the fetus is seriously and irreparably
damaged;
- when, during emergency action, an abortion is needed to save the
life of the woman.
Moreover, in practice, the procedural requirements to certify these
conditions have made the law more restrictive than it appears on its face
(Nowicka 1997, Penn
2001).
Yet our analysis suggests that in Poland, unlike in the United States,
public support for gender equality is a significant, but weak, source of
abortion opinions. The data in Table 1 suggest
that despite the efforts of such activist groups as the Polish Federation
for Woman and Family Planning, reproductive freedom as an important
component of gender equality has not succeeded as a plausible pro-choice
frame for the mass public. In contrast to the United States, Polish mass
publics are relatively unlikely to make a cognitive connection between
abortion and women's rights.
CONCLUSION
Our findings are obviously quite tentative and should be interpreted
with caution. In particular, we remind readers that the changes reported
in abortion attitudes in Poland and the United States are of limited
magnitude, and aggregate abortion attitudes remain stable in both
countries. Nevertheless, we observe visible, significant pro-life trends
among the mass publics in both the United States and Poland. These occur
during a period when each country has been among a handful of the
world's nations to limit abortion rights.
The two nations under consideration, in which legislators have placed
significant restrictions on access to legal abortion, exhibit interesting
similarities and differences. In both nations, we observe strong pro-life
period effects, which are partially offset by structural and compositional
changes in individual attitudes toward legal abortion.
In Poland, we observe two general trends: 1) a secular, pro-choice
trend, in which a more educated, secular population makes the development
of pro-choice attitudes more likely; and 2) a countervailing restructuring
of public opinion in which changing relationships between certain
predictor variables and abortion attitudes occasion strong pro-life
forces.
By contrast, there are small compositional changes, and more
pronounced structural changes, in the U.S. population, which would
normally be expected to increase support for legal abortion. However, the
effects of such demographic and attitudinal changes are dwarfed by the
existence of strong, profound, pro-life period effects. In our estimation,
the consistent period effects observed in the United States appear to be
relatively uniform across social groups, and seem likely to be
attributable to changes in the public framing of the pro-life case. In the
United States, abortion opinion seems responsive to general changes in
political discourse, which may have been initiated by strategic changes in
pro-life rhetoric by politically and religiously conservative elites.
In both countries, reduced access to legal abortion has important
consequences for women's health and life choices. This is especially
true in Poland, where abortion restrictions are relatively severe. In the
United States, legal restrictions are less onerous, but poor women do not
receive government subsidies for abortions (Harris v. McRae [1980];
Maher v. Roe
[1977]), and in many states, access to abortion providers
is limited. Moreover, state laws require doctors to counsel women on
abortion choices in restrictive ways, in some cases even requiring doctors
to provide medically discredited information, such as putative links
between abortions and breast cancer.
Yet there is relatively little discussion of how abortion affects
women's lives in either country. In both the United States and
Poland, the abortion issue has been effectively degendered. In Poland,
issues of national identity and religiosity have exclusively framed the
debate; in the United States, a pro-life frame has marginalized women in
policy discussions about abortion. It is worth noting, of course, that in
the United States, women activists on both sides of the abortion issue
offer competing visions of the effect of abortion on women's lives,
and this activist discourse is reflected in our data. In Poland, debate
about abortion is generally centered on the larger issue of the role of
the Catholic Church in Polish politics, and our results show a
correspondingly weaker relationship between gender role attitudes and
public responses to legal abortion. While net change in abortion opinion
in Poland and the United States seem quite similar, the manner in which
such modest changes come about is somewhat different in each country.
These similar outcomes can even come about through distinct processes in
different cultural settings. In each case, however, the ramifications for
public policy change are clear. Where public opinion can be shifted,
across time, by linking abortion to issues other than women and gender,
governments will find it easier to respond to demands to restrict
women's access to abortion.
Appendix: Summary of Measures and
Operationalization
Dependent variable—sum of pro-choice responses from ABHLTH,
ABRAPE, ABDEFECT, ABPOOR, ABSINGLE, ABNOMORE. If missing on two or fewer,
missing data is coded to .5. If missing on 3 or more, case is
excluded.
