Government programs and services “make up a basic and persistent presence in the lives of modern citizens … [providing] the basis for experiences of government-in-action far more regularly than do the activities that political scientists more commonly study, such as voting, contacting public officials, and participating in protests or demonstrations.”Footnote 1 Yet political scientists pay surprisingly little attention to what citizens know about them. Measures of political knowledge are typically limited to the names of prominent politicians and their roles, the positions taken by candidates and political parties, and the rules governing political institutions. Practical questions about benefits and services that are directly relevant to citizens' daily lives are rarely asked. This is an unfortunate omission. As students of policy feedback have observed, public programs can be politically empowering: they can shape perceptions of what governments ought to be doing, provide incentives for political mobilization, and furnish political resources that potentially compensate for socioeconomic disadvantage.Footnote 2 However, these positive effects cannot manifest themselves unless the intended beneficiaries of the various policies and programs are aware of their existence and know how to access them.
This type of knowledge may be particularly consequential for women given that they are more likely to be beneficiaries of public services and welfare state policies and to be employed in the public sector. Women's practical knowledge about government services and programs could potentially offset their disadvantage when it comes to more conventional forms of political knowledge. One of the most persistent and perplexing puzzles in the literature on political knowledge, at least as typically measured, is why women know less about politics than men do. This gender gap in political knowledge cannot be explained by differences in women's educational attainment or material resources or by the greater demands that child care responsibilities continue to make on many women's time. The gap has persisted despite the massive influx of women into the paid work force. Rich or poor, married or single, young or old, women tend to know less about politics than their male counterparts. This pattern holds in most established Western democracies.Footnote 3
Feminist scholars have suggested that these findings reflect gender-biased measures that implicitly treat politics as synonymous with the traditional arenas of electoral and legislative politics.Footnote 4 According to their critique, such measures necessarily underestimate women's political knowledge because they fail to ask about those aspects of politics that impinge on people's daily lives. The feminist critics suggest that we should broaden our conception of political knowledge by also asking citizens whether they know how to access welfare and other government services that are essential to their own well-being and that of their families. Feminist critics are not alone in their concerns about the questions conventionally used to measure people's knowledge of politics. Other critics charge that these questions have an elite biasFootnote 5 and are too narrowly focused on the governmental and electoral contexts.Footnote 6
Their criticisms of the measurement and conceptualization of political knowledge are well taken, and their call for a more encompassing approach to measuring political knowledge meshes well with the need for students of policy feedback “to trace the processes and specify the mechanisms that link public policies to mass responses.”Footnote 7 Knowledge of government programs and services may prove to play an important mediating role in explaining policy feedback effects.
In order to advance this agenda, we address four key questions. First, is this type of practical political knowledge measurable? Second, how do women and men compare when it comes to knowledge of government services and programs? Do the gender gaps disappear or even reverse, as the feminist critique would suggest? Third, how is practical political knowledge distributed: Is it subject to fewer social biases than its conventional counterpart? And more importantly perhaps, are the citizens who most need this information the most likely to possess it? Finally, can this type of knowledge contribute to the explanation of politically consequential behavior? The data come from a survey of 1,689 respondents in Montreal and Toronto, Canada's two largest metropolitan areas, which included a unique and novel collection of questions that tap how knowledgeable people are about a variety of government services and benefits.
Gender and Political Knowledge
One striking finding in the political knowledge literature has been that women typically know less about politics than men do.Footnote 8 For example, one of the most comprehensive assessments of gender differences in political knowledge in the United States found that the average gap across 51 knowledge questions was almost 10 percentage points in men's favor. Fully three-quarters of the women scored below the male median.Footnote 9 Comparable gender gaps in political knowledge have appeared with remarkable consistency across both established and emerging democracies. The size of the gaps may differ from one country to another, but the pattern is clear: Men consistently score higher than women on measures of political knowledge.
The gender gap in political knowledge has defied explanation. In general, education and income are strongly correlated with political knowledge.Footnote 10 However, the gender gap typically remains significant, even in the face of controls for education and income.Footnote 11 At one time, women's lower levels of political knowledge were attributed to their traditional roles as wives and mothers and their confinement to the domestic sphere. However, women have moved into the paid work force in massive numbers, and yet the gender gap in political knowledge persists.Footnote 12 More recently, scholars have focused on the increased pressures on women's time as they attempt to juggle the demands of their working and family lives. Yet, even the “double day”, it seems, does not explain why women continue to be less aware than men of what is going on in public life.Footnote 13
Another possibility is that women and men differ in their propensity to guess the answers to political knowledge questions.Footnote 14 The cultural norm that politics is a man's world may make it harder for men to admit to being ignorant of basic political facts. Conversely, women may be more reluctant to respond unless they are confident of their answer. Either way, the gender gap in political knowledge will be overstated. However, the jury is still out on how much of the knowledge gap is attributable to gender differences in willingness, as opposed to ability, to answer.Footnote 15 And even allowing for systematic gender differences in the propensity to guess, a sizable gender gap remains to be explained.
Feminist scholars have taken a different tack. They have been critical of the ways political knowledge is conventionally conceptualized and measured. First, most of the questions that have been asked to tap political knowledge refer to male politicians. It is plausible that women are more likely to be aware of female politicians and so including questions about female politicians could close or even reverse the gender gap in political knowledge.Footnote 16
Second, and more importantly, feminist scholars have taken issue with the way political knowledge is conventionally conceptualized. Feminist political analysts and theorists have long criticized the narrow vision of politics that has traditionally characterized political science.Footnote 17 Whether based on their different situational experiences and positions as wives, mothers, or public sector employees, or because of intrinsic differences, women are mobilized by different issues and priorities than men,Footnote 18 and their political engagement often blurs the lines between what has traditionally been considered the public and private spheres.Footnote 19 In particular, women are more likely than men to support consumer protection and regulationFootnote 20 and to prioritize welfare state-related issues such as education and healthcare.Footnote 21
Why might this be the case? Typically, explanations have traced gender differences in political attitudes and values to differences in men's and women's relationship with the state. Frances Fox PivenFootnote 22 has highlighted women's shared stake in safeguarding welfare state programs. Indeed, she maintains that “the main opportunities for women to exercise power today inhere precisely in their ‘dependent’ relationships with the state”.Footnote 23 Women's relationship with the welfare state is not confined to their role as clients; women are also more likely to rely on the welfare state for their employment. These differences in women's position vis-à-vis the state are reflected in overwhelming evidence that men and women differ in their policy preferences and priorities,Footnote 24 with women generally being much more supportive than men of an active role for government.Footnote 25 According to feminist critics, these differences in policy preferences and priorities may, in turn, affect the kinds of political knowledge that women possess.Footnote 26 Given women's greater reliance on the state and their greater stake in welfare services and programs, it seems plausible that women will possess more knowledge about government services, particularly when these are related to their daily activities and interests.
Marion SmileyFootnote 27 has picked up on this point. She suggests that conventional measures of political knowledge are gender-biased because they implicitly limit the scope of politics to the traditional arenas of electoral and legislative politics. As a result, her critique implies, these questions necessarily underestimate women's political knowledge. Knowing the names of prominent politicians, say, may capture aspects of political sophistication and general political interest, but no amount of political knowledge as conventionally conceived can help a woman to obtain welfare benefits or access the services that are essential to the well-being of herself and her family. Accordingly, Smiley argues, knowing how to obtain government benefits and access public services should be an important criterion for assessing a woman's political competence.
Highlighting the need to broaden our conception of politics when asking about women's political knowledge and assessing their competence as political agents does not mean that political knowledge as conventionally measured is unimportant for women. On the contrary, women need to be acquainted with basic facts about the political process and current politics. After all, it is in the traditional political arenas that decisions about what benefits and services are provided and who qualifies to receive them are made. However, this does not negate the potential importance of expanding our conception of political knowledge to encompass aspects that touch more directly on women's daily lives and concerns and speak to their policy preferences and priorities.
Indeed, government benefits and services are not just a female matter. Women in Western democracies may be the major beneficiaries of welfare state policies, but men benefit too—particularly in their role as parents and caregivers. Accordingly, a more general case can be made for adding knowledge of government benefits and services to our batteries of questions on political knowledge and civic competence. In a similar vein, Pippa Norris has underlined the importance of looking beyond the traditional political arenas to recognize that in their daily lives both women and men “may need practical information about the health risks from contaminated foods, their welfare rights at benefits offices, or how they can claim legal redress for grievances.”Footnote 28 Looking beyond knowledge of the traditional political arenas to knowledge of government programs and services becomes all the more important if we recognize the latter's potential for politically empowering feedback effects.
There has been remarkably little study of the public's knowledge of government services and benefits even though service provision constitutes a significant portion of what governments do. As a result, we have little understanding of how this type of political knowledge relates to its conventional counterpart and how such knowledge is distributed across various groups of the population. If feminist scholars are correct, this type of practical knowledge should not be subject to the same social biases as conventional political knowledge and so knowledge gaps defined by gender, education, income, racial background, and country of birth should shrink or even reverse when we look at knowledge about government services and benefits. It is people with low incomes who need to know what to do if they get a rent increase that is too high, minorities who need to know where to turn if they experience discrimination on the job or in housing, immigrants who have to cope with the challenges of settling into an unfamiliar environment, and so on.
Should we find that the gap fails to narrow, there are important implications for policy makers and service providers who need to ensure that the necessary information is reaching the target recipients. One of the very few studies in political science that examined people's knowledge of a government program found that low-income immigrants and people with very low incomes were the least likely to know about the federal Earned Income Tax Credit, even though they were among its target beneficiaries.Footnote 29 There is also some scattered evidence from the field of public health to suggest that information about government services and benefits is not necessarily reaching the potential beneficiaries. Families of children with chronic medical conditions are often unaware of welfare reform program provisions that entitle them to work exemptions,Footnote 30 and many parents of children potentially eligible for Medicaid are misinformed about eligibility rules.Footnote 31 Our research will be able to evaluate whether these findings hold for a wider array of government services and benefits.
Identifying differences in this type of political knowledge is important for another reason. Political knowledge translates into political power: Governments are more responsive to citizens' demands when political knowledge is evenly distributed across social groups.Footnote 32 There is no reason why this should apply only to conventional forms of political knowledge. Governments have more incentive to be responsive to citizens' needs for various types of services when they can be held accountable, and the people can only hold governments accountable for their actions when they know what governments are actually doing. Similarly, government programs and services can only be a source of political resources that compensate for socioeconomic disadvantage if the target beneficiaries are aware of their existence and how to access them.
This line of thinking suggests that practical political knowledge will have powerful political consequences. We predict that people who are aware of what governments are actually providing by way of a social safety net, protective services, and redistributive policies will be more likely to vote for candidates and parties that are committed to the maintenance or enhancement of such programs. We have no reason, on the other hand, to expect that conventional political knowledge will directly affect people's choice of party, though it may affect the bases on which they choose a party.Footnote 33 As Gilens observes, “Although knowing which office Ted Kennedy holds or how many justices sit on the Supreme Court does distinguish the politically knowledgeable from those who are less informed, and respondents who score high in such knowledge do express political preferences different from those of otherwise similar respondents who score low, this knowledge per se is unlikely to influence political choices or preferences.”Footnote 34 This type of political information is typically associated with greater ideological polarization and therefore does not significantly impact candidate choice.Footnote 35
Research Questions
We thus have four key research questions. First, we need to ask whether knowledge of government programs and services is measurable. This is not a straightforward task. The needs for practical knowledge about government benefits and services are not equally distributed. For example, knowing about maternity and parental leave benefits is only useful for parents or potential parents of child-bearing age. Similarly, knowing about services for the elderly may only be important for adults who have parents over a certain age. Questions have to be designed that relate to services and benefits that are potentially relevant to most citizens.
Our second research question focuses on the gender gap. The feminist critique implies that women will fare better than men on measures of political knowledge that tap more directly into women's everyday political worlds and their role as caregivers. Accordingly, our second task is to examine whether there are gender differences in knowledge of government benefits and services and whether women do, indeed, outperform men on these questions.
Our third research question switches the focus away from the gender gap in practical knowledge between men and women to address possible inequalities in the distribution of this form of knowledge among women. We need to determine what sorts of women know more—or less—about these practical matters. Based on their study of various dimensions of political knowledge, Delli Carpini and Keeter conclude that most citizens are political generalists rather than political specialists: “people who know a lot about one aspect of national politics also know a lot about others.”Footnote 36 Does this generalization hold for knowledge of government programs as well? Do the women who typically score high on conventional political knowledge score similarly well on this type of practical political knowledge, too? In other words, do women who are older, affluent, well-educated, and white come out ahead of other women, regardless of the type of political knowledge being measured? If the same women prove to have an advantage on both types of knowledge, there are important theoretical and practical implications. First, it would suggest that there is little analytical leverage to be gained by expanding our conception of political knowledge to include knowledge of government benefits and programs. From a theoretical perspective, asking about this type of knowledge would be redundant. Second, it would imply that the women who are most in need of government services and benefits are the least likely to know about them. Third, it could help to explain socio-demographic differences in the policy feedback effects of government programs and services: Socio-demographic biases in the distribution of practical political knowledge could limit the voices that governments hear on redistributive and welfare state matters.Footnote 37
Our final research question asks whether this type of knowledge actually does have political consequences. Does it contribute to the explanation of politically consequential behavior? Given the importance of women's views about social welfare in their vote choice in the US, UK, and Canada alike,Footnote 38 we expect knowledge of government services and benefits to be related to women's choice of party. The more aware women are of what services and benefits governments actually provide, the more likely they are to vote for parties that are committed to maintaining a strong social safety net, which generally means parties of the left.Footnote 39
Data and Methodology
Our data come from a 27-minute telephone interview with 1,286 women and 403 men randomly selected from 42 neighborhoods in Montreal and Toronto.Footnote 40 The response rate was 52 percent.Footnote 41 The survey oversampled women because its main purpose was to understand variations in women's political resources. However, a small group of men was included in order to be able to gauge possible gender gaps in the various resources. The neighborhoods were stratified by ethno-racial diversity, percentage foreign-born, and income in order to ensure that the sample included sufficient numbers of racial minorities, immigrants, and low-income respondents.Footnote 42 Montreal and Toronto were chosen because they are Canada's two largest major metropolitan areas and, as such, are home to many minority, immigrant, and low-income women, thus ensuring sufficient numbers to address our third research question regarding differences among women. According to Statistics Canada, 95 percent of all minority women, 88 percent of all female immigrants, and 66 percent of all Canadian females living in poverty reside in the 15 largest census metropolitan areas (with populations above 300,000).Footnote 43 Overall, the vast majority of Canadians live in urban areas. In fact, in 2001, 80 percent of all women lived in an area classified as urban by Statistics Canada, and 64 percent lived in an urban area with a population of at least 100,000.
The survey included both conventional political knowledge questions and questions about government services and programs (see Appendix for complete question wording). All but two of the questions used a multiple-choice format. This format has two advantages. First, it avoids the serious problems with coding open-ended knowledge questions that have recently become apparent with the ANES questions measuring political knowledge.Footnote 44 Second, it minimizes the possible impact of respondents' differential propensity to guess.Footnote 45 Alternative formats, such as true/false and either/or, make it easier to guess the answer which may lower the reliability (and hence, the validity) of itemsFootnote 46 and cause gender gaps in conventional political knowledge to be overstated. Mondak recommends that “don't know” responses be actively discouraged.Footnote 47 However, based on their split-ballot CATI experiment, Sturgis and his colleagues have questioned the wisdom of this approach.Footnote 48 Accordingly, we neither encouraged nor discouraged “don't know” responses.Footnote 49
The conventional knowledge questions focused on all three levels of government in Canada, thus avoiding the “national fixation”.Footnote 50 Respondents were asked for the names of Canada's governor general (who is a woman), the prime minister, the official opposition party in the House of Commons, a female cabinet minister, the judge who headed an inquiry into a major political scandal, the premier of their province, and the city's mayor. The number of correct answers to these seven knowledge items comprised the conventional political knowledge scale (Cronbach's alpha: .78).
The questions about government services and benefits focused on those that are potentially relevant to most citizens and also spanned all three levels of government. The questions dealt with housing, health, law and protection, and tax benefits. Renters were asked about the best place to go if someone wanted to contest a rent increase, while home owners were asked a parallel question about obtaining a building permit to make renovations to their residence. The health questions asked about the cost of screening tests. Since the type of test that is routinely administered varies by age and sex, parallel forms were used. Women under 50 were asked whether Pap tests are free under their provincial health plan or cost $20, $50, or $80. The parallel question for men in this age group asked about tests for sexually transmitted diseases. Women aged 50 and over were asked a parallel question about mammograms, while their male counterparts were asked about screening for prostate cancer. Three questions focused on issues relating to law and protection. First, respondents were presented with a hypothetical scenario: “If someone was refused an apartment and thought it was because of their racial background, where would be the best place to make a complaint?” Awareness of legal aid was tapped by asking about the best place to go if someone had to go to court and could not afford a lawyer. Respondents were also asked where to go if they knew of a child being abused. Finally, we probed awareness of federal tax benefits and credits by including two yes/no items about the Goods and Services Tax (GST) credit and the child tax benefit.Footnote 51 Responses to all seven questions were combined to create an additive scale (Cronbach's alpha: .55).
We begin by presenting the female-male differences on each of the knowledge questions. We then estimate separate regression models for each of the two knowledge scales. This enables us to determine the size, sign, and significance of the effects of gender on the two types of knowledge, controlling for other social background characteristics. The independent variables were selected on the basis of the literature on political knowledge and include key correlates of conventional political knowledge such as age, education, income, employment status, and racial background, as well as immigrant status.Footnote 52 We also controlled for city and mother tongue and for other variables such as having children and being married. All of the control variables were entered as dummy variables, with the named category coded one. In order to test whether the effect of gender is significantly different for the two types of knowledge, we created a stacked data matrix in which each respondent appears twice. Such models are commonly used in electoral research.Footnote 53 In the first stack, the respondent's knowledge score corresponds to the conventional political knowledge scale while the second stack uses the respondent's score on the practical political knowledge scale. Each independent variable is interacted with a stack variable that is coded zero for the conventional knowledge stack and one for the practical knowledge stack. This specification enables us to determine whether the effects of sex differ significantly for the two types of knowledge, even controlling for possible differences in women's and men's social background characteristics. If the effect of sex is indeed different for the two types of knowledge, the corresponding interaction terms will be statistically significant.Footnote 54 Because the same respondents appear twice in the stacked data set, the analysis violates the assumption of independence. To correct for the possible inflation of statistical significance, we have clustered by respondent and performed the estimates using robust standard errors.Footnote 55 Since the dependent variables are both scales ranging from zero to seven, the models are all estimated using ordinary least squares regression.
In order to answer the question relating to the social distribution of the two types of knowledge, we re-estimate the regression models for women only (see table 2). Comparing the social correlates of the two types of knowledge will tell us whether knowledge of government services and benefits is more evenly distributed across different groups of women than its conventional counterpart. This comparison will also enable us to assess whether this more practical type of political knowledge is empirically and not just conceptually distinct.
In the final set of models, practical political knowledge becomes the independent variable and its effects on women's intended vote choice are examined (see the (see appendix for question wording). The model is only estimated for Toronto women. Since the advent of the Bloc Québécois in the early 1990s, elections in Quebec have revolved to a remarkable degree around the issue of Quebec sovereignty, as opposed to issues relating to health and other forms of social spending that typically dominate general elections in the rest of the country.Footnote 56 As such, there was no reason to expect that knowledge of government services and benefits would affect vote choice in Montreal.Footnote 57 Moreover, the Bloc only runs candidates in Quebec, making it problematic to estimate a single model for Montreal and Toronto respondents combined. The dependent variable has three categories: Conservative, Liberal, and New Democratic Party or Green, corresponding to a vote for a party of the right, a party of the center and a party of the left, respectively.Footnote 58 Estimation is by multinomial logistic regression. Controls are included for social background characteristics, as well as conventional political knowledge.
Findings: Gender Gaps in Political Knowledge
The first question we need to answer is whether gender gaps in political knowledge disappears or even reverses when it comes to knowledge of government services and benefits, as some feminist scholars would predict. In order to compare the gender gaps in conventional political knowledge and practical political knowledge, we have subtracted the percentage of men answering a given item correctly from the percentage of women giving the correct answer. The female-male differences are presented in figure 1, along with their statistical significance.
Gender differences for types of political knowledge
Bars represent gender gaps for various political knowledge items, with statistical significance attached, *P < .10; **P < .05; ***P < .01
For conventional knowledge, all of the statistically significant gender gaps are in men's favor. Only two items fail to produce statistically significant differences: the prime minister's name and the name of the governor-general (a largely ceremonial office). The latter is a woman, which might seem to indicate that gender differences disappear when the office-holder is female. However, surprisingly few women could name a female cabinet minister, even though the multiple-choice format required only name recognition as opposed to recall, and the fact that men were significantly more likely than women to correctly identify a female cabinet minister casts doubt on the notion that including more female politicians in knowledge batteries will necessarily close the knowledge gap.
The size of the gap clearly depends on the question asked: Easier questions tend to elicit smaller gaps (see table 1). For example, like their male counterparts, most women in this sample knew the prime minister's name, and most knew the name of their mayor (see table 1). However, the numbers fell and the gaps widened when women were asked the name of the judge who headed the commission of inquiry into the sponsorship scandal (despite the scandal's prominence in the news) and the party forming the official opposition. Overall, about 25 percent of the women answered all seven of the conventional political knowledge items correctly, compared with 35 percent of men.
Percentage answering knowledge questions correctly
In the case of government services and programs, by contrast, none of the statistically significant gaps favor men. Women scored as high as men or higher on all seven questions. On four of the items there is no difference, and on three of the items the gender gap actually reverses to favor women. Women were significantly more likely than men to answer the health question correctly, which is consistent with the result reported by Pippa Norris for European women's awareness of the risks of skin cancer.Footnote 59,Footnote 60 Women were also significantly better informed about where to turn if they knew of a child being abused or if someone experienced racial discrimination. However, from a policy point of view, the numbers are hardly a cause for complacency. For example, a surprisingly high number of women did not know that screening tests for cervical cancer and breast cancer are free of charge under their provincial health care plans. This underlines the importance of determining whether the women who most need to know about particular services and programs are the most likely to have that information, a question to which we will return.
Overall, 26 percent of the women knew the answers to all seven practical questions, compared with only 13 percent of the men. It might be argued, though, that women are scoring higher than men on some of these practical knowledge items because the questions have a female bias. In other words, men may be disadvantaged because the questions deal with domains that are traditionally seen as being female. For example, because women are often the primary caregivers, they might simply be more aware about child abuse, which would explain why they are more likely to know where to turn if they encountered an incident of child abuse. However, the health question asked men about screening tests for prostate cancer or STDs (depending on their age), and the question about the human rights commission referred explicitly to racial discrimination, which a priori is not gendered (though it may intersect with gender discrimination). Thus, it is unlikely that the observed gender gaps reflect some inherent female advantage in the topics addressed.
At first glance then, the feminist critique has been validated: Focusing solely on conventional political knowledge understates how much women know. While the conventional items produce the classic gender gap, all of the items involving more practical matters cause the gap to disappear. To the extent that either sex has the advantage, women know more about basic benefits and services. The question now is whether this pattern holds in the face of controls for social background characteristics. Do women continue to perform as well as (or better than) men on these practical political knowledge questions once other social background characteristics, such as education and marital status, that might affect their knowledge about social programs and benefits are taken into account?
It turns out that even controlling for a variety of social background characteristics, women continue to score significantly lower than men on the conventional political knowledge scale and significantly higher than men on the government services and benefits scale (see columns one and two in table 2). Moreover, the difference between the two gender coefficients is statistically significant. This can be seen in column three, which presents the coefficients for the interaction terms from the stacked analysis (see above).Footnote 61 The coefficient of concern here is the one for gender which is significant at the p < .001 level, confirming the reversal of the gender gap between the two knowledge scales. Note, though, that in absolute terms, the gender gap is smaller for knowledge of government services and benefits. Indeed, when health is dropped from the scale, the gender coefficient for this type of knowledge no longer achieves statistical significance (supplementary table 1). The most prudent conclusion, then, is that men outperform women when it comes to conventional political knowledge, but women and men generally have similar levels of knowledge about government programs and services.
The impact of sex by type of political knowledge
Note: Columns one and two show the coefficients and standard errors from an ordinary least squares regression using the conventional and practical knowledge scales, respectively, as the dependent variables. The entries in the third column are the coefficients for the interaction terms derived from a stacked analysis where each social background characteristic is interacted with a stack variable that is coded zero for conventional knowledge and one for knowledge of services and benefits. Estimation is again by ordinary least squares regression. The stacked analysis was clustered by respondent to correct for non-independence.
*** p < .01;
** p < .05;
* p < .10
Disparities in Women's Political Knowledge
Once we broaden our conception of political knowledge to consider practical knowledge of government services and benefits, women prove to be at least as well informed as men. However, disparities among women should be of as much concern as—if not more concern than—differences between women and men. We need to know whether social disadvantage makes for smaller stocks of knowledge among women. This is a question with important theoretical and policy implications. If the same types of women score well on both conventional and practical knowledge, there is not much point in expanding our conception of political knowledge to include the latter. Meanwhile, from a policy perspective government services and programs are not being delivered effectively if the women who are most likely to need them are unaware of their existence or unsure how to access them.
It is apparent that practical political knowledge is more evenly distributed across social lines than conventional political knowledge (see table 3). Notably, the education-related disparities that are so evident for conventional political knowledge effectively disappear for knowledge of government benefits and services. College educated women have significantly higher scores on the conventional political knowledge scale, while women with less than a high school education have significantly lower scores. The corresponding effects for knowledge of government benefits and services are too weak to attain conventional levels of statistical significance. Moreover, the difference between the education coefficients for the two types of knowledge is statistically significant. Age follows a similar pattern. Younger women score significantly lower on the conventional political knowledge scale, but the gap narrows to insignificance when it comes to the acquisition of more practical political knowledge, and again the difference is significant. The same is true of women with children under 18 at home. The presence of children may be an impediment when it comes to keeping abreast of conventional politics, but not when it comes to acquiring information about government services and benefits. Neither employment nor marital status have any effect at all on conventional political knowledge, but they both give a modest boost to knowledge of government services and benefits, attesting to the importance of situational factors for this type of knowledge.
The impact of social background characteristics by type of political knowledge (Women only)
Note: Columns one and two show the coefficients and standard errors from an ordinary least squares regression using the conventional and practical knowledge scales, respectively, as the dependent variables. The entries in the third column are the coefficients for the interaction terms derived from a stacked analysis where each social background characteristic is interacted with a stack variable that is coded zero for conventional knowledge and one for knowledge of services and benefits. Estimation is again by ordinary least squares regression. The stacked analysis was clustered by respondent to correct for non-independence.
*** p < .01;
** p < .05;
* p < .10
These results confirm that expanding our notion of what constitutes political knowledge extends the circle of who qualifies as being knowledgeable to include not only more women, but women who are younger or have less education. Moreover, having children is not a disadvantage for women when it comes to knowledge of government programs. The linguistic and city differences disappear as well. Clearly, these knowledge questions are tapping a different kind of political knowledge, one that is distributed more evenly across society. As such, this type of knowledge deserves further conceptual attention.
At the same time, though, from a policy standpoint, the results give cause for concern. Whichever type of knowledge we consider, there is a pronounced gap between immigrant women and Canadian-born women. This is especially true of women who arrived within the past ten years. The knowledge gap between recent immigrant women and other women is more than 40 points when it comes to knowing where to take a complaint related to rented housing and 34 points for legal aid (supplementary table 2). However, even women who have been in the country much longer remain significantly less knowledgeable than Canadian-born women about a variety of government programs and services. The significant negative coefficients for immigrants and recent immigrants in table 3 for both types of knowledge speak to these knowledge deficits. Women who are new to the country face enormous challenges in settling into their new environment. More than any other group of women, they need to have basic information about the availability of government services and benefits, but this vital information is clearly failing to reach many of these women.
The second disparity is defined by income. Low-income women score significantly lower than other women on both types of knowledge. In fact, there is a 20- to 30-point difference between the lowest and highest income groups on nearly all of the questions about government services and benefits (supplementary table 3). The lack of significant interaction effects in table 3 confirms that low-income women are as disadvantaged when it comes to knowledge of government benefits and services as they are when it comes to its conventional political knowledge. Similarly, affluent women are just as advantaged on the two types of political knowledge.
Finally, older women are significantly better informed than younger women about politics as conventionally measured by political knowledge questions. However, the gap reverses for the other type of political knowledge where they are at a significant disadvantage. Like immigrant women and low-income women, older women are likely to have greater need of government services and benefits and yet many of them appear to be unaware of what is available.
Still, overall, knowledge about government benefits and services is more evenly spread among women than its conventional counterpart. Impediments to conventional political knowledge such as having children or being young do not exert the same negative effects on this type of practical political knowledge. However, the analysis has also alerted us to the knowledge gaps for immigrants, particularly recent arrivals, and for low-income women that remain in place even when political knowledge is defined more broadly. These gaps are even more striking when immigrant status and poverty intersect.Footnote 62 Immigrant women with household incomes in the lowest third of the sample distribution score fully one point lower (4.2) on the practical political knowledge scale than Canadian-born women with similarly low incomes (5.3); they also score one point lower than other immigrant women (5.2).
Vote Intention and Political Knowledge
The expectation was that women who are knowledgeable about government services and benefits will be more likely to vote for a party of the left that supports a strong social safety net (the New Democratic Party or the Green Party) and less likely to vote for a party of the right (the Conservative Party) than people who know less about what governments actually do.Footnote 63 This prediction is clearly supported by the data. The more people know about various government programs, the more likely they are to say that they would vote for the NDP or the Green Party over the Conservative Party.Footnote 64 As predicted, conventional political knowledge does not significantly affect women's choice of party.
Types of political knowledge and vote intention (Toronto voters)
Note: The column entries are logistic regression coefficients with robust standard errors shown in parentheses. Estimation is by multinomial logistic regression.
*** p < .01;
** p < .05;
* p < .10
The effect of knowledge of government services and benefits is not only statistically robust, but fairly substantial. For example, the estimated probability of a Conservative vote intention is 15 percentage points lower for a voter who answered all seven questions correctly, compared with a voter who only got three answers right (see figure 2). The corresponding estimated increase in NDP or Green Party vote intentions is 18 percentage points. Given the closeness of recent general elections in Canada, the effect is also politically consequential: The 2004, 2006, and 2008 elections all resulted in minority governments.
The impact of knowledge of government services and benefits on voting for the left versus the right
Conclusion
Feminist critics and other critics have called for a broadening of our conception of political knowledge. Their unease with existing approaches finds echoes in Norris' objection that conventional political knowledge questions are too narrowly focused on traditional political arenas.Footnote 65 At the same time, critical voices have expressed dissatisfaction about the inability of our field to bridge the study of public policy and mass politics. In this article, we have taken these critiques seriously by adopting a more expansive view of what constitutes political knowledge and by asking questions that relate more directly to people's everyday lives and concerns. Using a novel series of practical political knowledge questions, some of the expectations of the feminist critics have been confirmed. We find that the gender gap closes (and for some items, even reverses) when men and women are queried about government benefits and services. However, this is perhaps a less female-specific form of knowledge than some feminist accounts would lead us to believe. When we controlled for social background characteristics, women still scored significantly higher than men, but this result was largely driven by the health item and ceased to be statistically significant once that item was removed. Moreover, we find only mixed support for the argument that women will fare better on conventional political knowledge items that involve female politicians.
The fact that the knowledge gap between women and men closes when we adopt a more expansive conception of political knowledge is hardly cause for satisfaction given that we go on to find even larger gaps among women. Practical knowledge is more equally distributed among women than its conventional counterpart. We have, for example, shown that information about government benefits and services reaches groups of women, such as those with less formal education, the young, and people with children, who tend to fare poorly on measures of conventional political knowledge. However, while some of the social biases associated with conventional political knowledge are diminished or disappear altogether when we look at knowledge of government services and benefits, there are two glaring exceptions—low-income women and immigrant women. Along with older women, these women are the most in need of a strong social safety net, and yet they are the least likely to know about various government benefits and programs. There is an important implication here for public policy. Clearly, governments and service providers need to do much more to ensure that information about their programs and services reaches the people who need it most.
At the same time, though, the fact that socio-demographic differences are less pronounced and that some differences disappear altogether suggests that our measures do indeed capture an important new dimension of political knowledge, one that is conceptually and empirically distinct from conventional political knowledge.
This type of knowledge has political consequences. The more women know about these matters, the more likely they are to vote for a party of the left and the less likely they are to be attracted to a party of the right. Women's awareness of what governments are—or are not—doing to maintain and strengthen the social safety net clearly translates into a higher vote for the left. One implication is that parties of the left can potentially increase their share of the female vote by crafting campaigns designed to increase women's knowledge of government services and benefits. A second implication is that the gender gap in vote choice might be wider if more women possessed this type of practical knowledge.
This is not to downplay the importance of conventional political knowledge. Women need to have some awareness of the traditional political arena, given its role in determining such fundamental matters as who is entitled to welfare assistance and how much should be allocated to its provision. However remote this arena may seem to a woman who is struggling to make ends meet, what transpires there is of crucial importance to her ability to feed and house her family. Women who possess a modicum of information about politics will be better able to voice their political needs and wants and thus to influence political decision making. Items of conventional political knowledge may often resemble trivia or textbook knowledge, but they do provide a good indication of people's general level of political awareness.Footnote 66 However, they provide only a partial picture of what citizens need to know. Conventional political knowledge may be helpful when it comes to matching policy preferences with the correct party,Footnote 67 but knowing the length of a president's term in office or the name of his national security advisor is not much help when it comes to seeking redress for discriminatory treatment or saving a child from an abusive parent. A realistic assessment of citizens' political knowledge needs to focus on what they need to know, and that includes practical knowledge of government benefits and services.
Broadening the measurement of political knowledge to include citizens' knowledge of what governments are actually doing also fits into the agenda of students of policy feedback. As Suzanne Mettler and Joe Soss observe, “The consequences of public policies for mass political behavior are a subject ripe for investigation.”Footnote 68 The use of practical political knowledge measures that tap into citizens' awareness of government programs and services constitutes an important missing element in the study of the link between public policy and political behavior. First, measures of this type of knowledge can help to illuminate why government programs vary in their feedback effects by identifying which programs are widely visible and which are more hidden from public view. Second, they can provide insight into the processes that underpin policy feedback effects as citizens' awareness of programs will surely prove to be a critical mechanism linking public policies to mass political behavior. Finally, measures of practical political knowledge can help us to understand why the feedback effects of public policies vary across individuals and groups.
There are several other directions that future research on practical political knowledge should take. We focus here on three: developing more measures of practical political knowledge; examining the effects of such knowledge on other forms of political behavior; and investigating the importance of this type of political knowledge for other social groups.
First of all, the indicators of practical political knowledge need to be expanded. Comparable indicators need to be developed across different welfare state contexts in order to assess their cross-national validity. At the same time, making distinctions among different types of practical political knowledge might prove theoretically and empirically valuable. For example, knowledge measures that distinguish between universal and targeted programs could reveal significant differences in the political usefulness of practical political knowledge. So could measures that differentiate between programs that benefit the respondent and those that mostly benefit fellow citizens.
At the same time, future research needs to analyze the effects of practical political knowledge on a wider variety of political behaviors. So far we have been able to show that practical political knowledge matters to vote choice. It may also have a mobilizing effect on various forms of political participation. It is likely to be particularly useful for those political acts that are intended to change public policy and patterns of service provision by targeting government officials, institutions, or political parties. However, scholars of political participation only rarely ask questions about the targets and nature of participatory acts, and so we lack a clear understanding of the role played by particular types of political knowledge in explaining various ways of being politically active.
Finally, our research was initially inspired by the puzzling and persistent gap in political knowledge between men and women and by feminist critiques of traditional political knowledge measures. However, as noted above, several other critics have noted the need to expand the definition and measurement of political knowledge. So, the question becomes whether our findings can be transferred to other groups as well. As men are becoming more reliant on the welfare state, it will be particularly interesting to see whether their knowledge about government programs and services increases and has the same political consequences as we found for women.Footnote 69
This type of political knowledge taps an important dimension of democratic performance, namely the link between voting and policy responsiveness. If citizens are generally unaware of what their governments are actually providing in the way of programs and services, they will be less likely to match their preferences with their vote choice. Practical political knowledge about government services and programs thus becomes a critical resource not just for women but for all citizens.
Appendix
Conventional Knowledge Scale—Question Wording (correct answers are with an asterix)
(1) We would like to know how widely known some political figures are. [Toronto] Is the current Premier of Ontario Mike Harris, Dalton McGuinty*, John Tory or Howard Hampton? [Montreal] Is the current premier of Quebec Bernard Landry, Jean Charest*, André Boisclair or Mario Dumont?
(2) Is the judge heading the inquiry into the sponsorship scandal Alan Gold, Irvin Cotler, John Gomery* or Beverley McLachlin?
(3) Do you happen to recall the name of the mayor of [Toronto] [Montreal]? [Toronto] Is it Mel Lastman, John Sewell, Hazel McCallion or David Miller*?/[Montreal] Is it Pierre Bourque, Jean Doré Gilles Vaillancourt or Gerald Tremblay*?
(4) Is the new Governor General of Canada Sheila Fraser, Michaëlle Jean*, Adrienne Clarkson or Jeanne Sauvé?
(5) And the name of a current female Cabinet Minister in Ottawa: Anne McLellan*, Kim Campell, Sheila Copps, or Deborah Gray?
(6) Do you happen to know which party is the official opposition party in Ottawa? Is it the Bloc Québécois, the Conservative Party of Canada, Liberal Party*, or NDP (New Democratic Party)?
(7) And who is the current Prime Minister of Canada: Stephen Harper*, Jean Chrétien, Paul Martin or George Bush?
Practical Knowledge Scale—Question Wording
(1) [Renters] If someone got a rent increase that was too high, where is the best place to go to get it reduced: City Hall, [the Rental Board] [Régie du Logement], the Ministry of Housing, or the Canada Mortgage and Housing Corporation? [Owners] If someone wanted to make renovations to their home, where would they go to get a building permit: the Real Estate Board, City Hall*, the Ministry of Housing, or the Canada Mortgage and Housing Corporation?
(2) [Women 50 and over] Under OHIP, is a mammogram free*, $20, $50 or $80? [Women under 50] Under OHIP, is a Pap smear test free*, $20, $50 or $80? [Men 50 and over] Under OHIP, is a prostate exam free*, $20, $50 or $80? [Men under 50] Under OHIP, is a test for sexually transmitted diseases (STDs) free*, $20, $50 or $80?
(3) If someone was refused an apartment and thought it was because of their racial background, where would be the best place to make a complaint? Is it the Ombudsman of [Ontario] [Quebec]; [Ministry of the Attorney General] [Department of Justice]; the Police; the [Ontario] [Quebec] Human Rights Commission*; or [the Rental Board] [Régie du Logement]?
(4) If you knew of a child being abused, where would be the BEST place to go? Is it the school board or school; Youth Justice Services; [Children's Aid*] [Director of Youth Protection*]; Ministry of Children and Youth; or the Police?
(5) If someone had to go to court and could not afford a lawyer, where would be the BEST place to go? Is it the Ombudsman of [Ontario] [Quebec]; Legal Aid*; Ministry of the Attorney General; or the [Ontario] [Quebec] Bar Association?
(6) People with low incomes can claim a variety of benefits when they file their federal tax returns. Can people with low incomes claim a GST tax credit? [yes*]
(7) Can they claim a Canada Child Tax Benefit? [yes*]
Vote Intention
The following question was used in order to measure vote intention.
If a federal election were held today which party would you vote for: Liberal, Conservative, NDP, or another party?
Liberal Party
Conservative Party
NDP
Green Party
Other, specify
Volunteers: wouldn't vote
Supplementary Materials
Explanatory file http://journals.cambridge.org/pps2010003
Summary Data http://journals.cambridge.org/pps2010004
