No abstract

Scholars have long suggested that population diversity should serve as a check on majority tyranny, but empirical evidence for this diversity hypothesis is inconclusive. This essay suggests that mixed findings arise because of the way in which scholars have translated Madisonian writings into statistical tests. It argues that the Federalist papers actually suggest an interactive relationship, where diversity moderates the effectiveness of representative institutions in protecting minority rights, rather than a direct relationship between population homogeneity and repression. The article tests for this moderating effect via a mixed methods approach, which includes a quantitative analysis of the passage of same-sex marriage bans in the American states, as well as a qualitative examination of legislative activity in four states.

Background Amazon Mechanical Turk (MTurk) is an online labor market in which people ("requesters") requiring the completion of small tasks ("Human Intelligence Tasks" [HITs]) are matched with people willing to do them ("workers"). MTurk has become a popular data collection tool among social science researchers: In 2015, the 300 most influential social science journals (with impact factors greater than 2.5, according to Thomson-Reuters InCites) published more than 500 articles that relied on MTurk data in full or in part (Chandler & Shapiro, 2016). Reflecting the popularity of MTurk, considerable effort has been invested in evaluating data collected from it, with particular emphasis on documenting the demographic and psychological characteristics of its population, the quality of respondent data, and the methodological limitations of the platform. As a result, MTurk workers have become one of the most thoroughly studied convenience samples currently available to researchers (for a review, see Chandler & Shapiro, 2016), and researchers have learned a great deal about the ways in which MTurk respondents are and are not similar to the general population. There are reasons to suspect, however, that there are also important variations between different samples drawn from MTurk, and these variations have received far less attention. This article addresses this question, using data from a study of approximately 10,000 MTurk workers to examine whether sample composition varies as a function of the time that it is collected. We begin by reviewing what extant research reveals about the demographic composition of the MTurk worker pool. Then, we describe the methods and measures that we use in our study, after which we present the results of our analyses, 712774S GOXXX10.

No abstract

No abstract