What was the Harell article's findings on voter turnout and associational involvement? (3)
1. Little to no gender gap at the federal or provincial level. But at municipal level, women of European decent participate more.
2. For first generation immigrants, there is more of a traditional gender gap, but this closes in the following generations, particularly for non-euro descent
- 3. Reverse gender gaps where women have more associational involvement (not including sports clubs), but this varies among different ethnic groups
- This can be seen as an indicator of political participation
What was the sampling challenge previous researchers faced in the Harell article?
Because of small sample sizes for minority groups, it has been difficult in attaining a representative sample of certain ethnic groups. Researchers required to collect oversamples among specific communities or seek out alternative data sources, such as association membership records or official voting statistics.
1. What was Harrel's study about?
2. What was the overall finding/pattern?
3. What were the two variables of focus in this article?
1. Situating different women’s experiences not only as women, but as members of various ethnocultural communities is important for understanding the ways in which gender influences political behaviour.
Attempts to find a place for intersectional feminist insights in the study for gendered political behaviour in Canada.
2. Ethnoracial and immigration cleavages appear to be much stronger dividing lines for electoral participation than gender.
3. Ethnicity and immigration status
1. What was the secondary data set Harrel drew from?
2. What was the issue she faced using this data set?
3. What is her critique of the previous literature with regards to methodological approaches for gender studies in Canada? (2)
- 1. The Ethnic Diversity Survey (2002)
- by Stats Canada which looked at general voting patterns at the federal level, as well as involvement in civic associations.
Had an exceptionally large sample of 41,695 participants.
2. There were no precise political attitude indicators, so she had to figure out how to operationalize this, and come up with something meaningful with the data that she has
3a. Previous methodologies apply a binary gender-based comparison which tells us little about differences between
women by ignoring other clevages such as class, race, ethnicity and sexual orientation.
3b. Gender studies is predominantly based on survey research. Most available data over-represents the majority population of white, middle-class, heterosexual women.
What are the 3 hypotheses Harell used?
1. Differences in ethnic communities should be explained by the large numbers of immigrants in those groups (as opposed to Canadian's with European backgrounds which will have fewer immigrants, and so fewer differences).
2. Gender gaps in minority ethnic communities are expected to be larger because of the challenges these women face to to lack of resources.
3. Differences between ethnic minority groups in political and civic participation should be explained by the different access to socio-economic resources
What are two challenges for a quantitative analysis for studies like Harrel's on intersectionality?
1. Comparing quantitative indicators of men and women relies on averages, which necessarily overrepresents the values, attitudes and behaviours of the more numerous, which are the white, middle class, heterosexual women (and men) who have historically made up the largest proportion of survey samples.
2. Inntersectionality is often associated with postmodern and poststructuralist approaches to politics that are critical of positivist approaches that use quantitative analysis.
What is the main purpose of Bittner and Goodyear-Grant's article?
What was their hypothesis?
To provide a more nuanced approach to attitudinal analyses by considering the role of salience of gender identity on political attitudes.
* By salience, they mean strength of attachment to the group
Hypothesis: We anticipate a scale from left to right, in which gender salient men are on one pole and gender salient women are on the other, with non-salient men and women in between them
What was the methodology used in Bittner and Goodyear-Grant's study? (3)
1. The conducted their own surveys (as opposed to Harrel) which were a series of web-based surveys conducted during provincial elections in Canada over 2011 and 2012, with 3 measures of gender (including a measure of gender group salience).
* Surveys were self-administered, so less susceptible to interviewer effect
2. The survey instrument included demographic variables, long-term political dispositions, attitudes about specific political issues, and perceptions of party leaders and economic conditions.
3. Feeling thermometer that asked how closely the participant identified with their gender group, from 0 to 100 (100 being extremely close)
What were the three parts of the analysis in Bittner and Goodyear-Grant's study?
1. Look at salience distribution to see whether voters are homogenous in salience to their gender identification.
2. Look for the sources of the variation in gender salience.
3 To look at the role of gender salience in political attitudes with respect to the gender gap
What were three findings in Bittner and Goodyear-Grant's study?
1. Over 40 per cent placed themselves at the very end of the scale, rating their gender salience at 100, while only about 10 per cent of respondents fall at the mid-point or lower.
2. While both women and men feel close to their gender group in nearly equal proportions, women are more likely to claim that their gender identification is not salient.
3. Gender salience higher among religious, married, older, and from Quebec and NL, and lower among university graduates
What were the conclusions of Bittner and Goodyear-Grant's study?
1. In almost every case, the gender gap is larger between women and men (and women are more to the left) among those in the high-salience category.
2. When we incorporate gender group salience the differences between women and men remain, but there are also more differences among women
3. Gender gaps in political attitudes mostly disappear when women have low levels of gender group identification,
What are the three aspects to consider when using big data as a sample? (Nagler and Tucker)
1. we have to consider how representative the sample is of the population that we claim to be studying
- 2. we have to consider whether the population from which the sample was drawn is a population
of interest and/or whether the population it was drawn from generalizes to other populations.
- 3. because big data is big, we have to adopt more caution than typically is the case when interpreting our measures of statistical precision.
* Big data drives sampling error toward zero but does not reduce other errors associated with inferences drawn from a sample.
What are three reasons social media data is valuable to political scientists?
1. it allows us to observe a rich set of data on those networks in which people choose to embed themselves
2. social media data allows us to observe the comments that people freely make about politics, so we can find out what people think
3. Because so many people are on social media, we can obtain large samples from subpopulations of interest often not available in traditional types of survey data
What is a problems with using Twitter to gather data?
People could claim to be someone they are not
How can Twitter lead to greater political participation? (4 things)
1. can provide people information about opportunities for participation, such as where a rally is located
2. can provide people information about the cost of participation. They can find out if the protest is well attended, if they are worried about a clampdown from a hostile regime for example.
3. can inform people about potential benefits of participation, through information about what policies a candidate stands for
4. can also provide information about what users’ friends think about politics, which can influence their political decisions
What are the inferences NOT to draw from Twitter? (2 points)
1. Cannot make a correlation between a number of tweets and votes for a candidate. This is a spurious relationship. The correlation is the money spent, the number of supporters in their riding etc.
2. Twitter is not the only source for political information, so we have to be careful about inferring that people were informed by Twitter alone.
To find out if Twitter use is having an impact on political behavior and politics, we would want to do the obvious: compare Twitter users to non-Twitter users.
What makes Twitter interesting for political science? (2 things)
1. Twitter data is a potential source of information about the political world, just as telephone and Internet surveys are
2. Twitter is a potential variable itself in the political world by providing politically relevant information to people that has the potential to alter their political behavior
What were the limitations of the previous studies that Bittner and Goodyear-Grant looked at, and were cautious to avoid?
Limitations to do with question wording. Studies that focused on salience in other categories such as partisanship did not include categories of independent or no party affiliation, which inflated the results for partisanship.