Dependent Variable. Given the hypothesis that CDF deployment increases the likelihood of COIN victory, COIN outcome is the dependent variable. Drawing on Stam (1996) and Xxxxx and Xxxxxx (2009), the variable can take on three discrete values: win, draw, or lose. Operationally, insurgencies end either as a result of a military defeat or by way of a negotiated settlement. In the event of a military outcome, an incumbent is considered to have lost if its armed forces are destroyed or evicted from the capital. Likewise, an incumbent is considered victorious if the insurgent organization is destroyed or substantially weakened. Whenever a decisive military outcome is not observed at an insurgency’s end, either a compromise settlement has been negotiated or one of the participants has conceded all rival demands. An insurgency is considered to have ended in a draw if the warring parties reached a compromise agreement in which the government conceded some, though not all, insurgent demands. The second possibility is that an insurgent gives up its armed struggle having received no government concessions or hostilities end because the incumbent concedes all or nearly all rebel demands. In such cases, the outcome is coded as a victory and defeat, respectively. In coding outcomes, I use a narrow definition of draws in that the insurgent movement must obtain either a power- sharing or autonomous arrangement. The possibility exists that a rebel organization’s initial aim is limited to a power-sharing or autonomous arrangement, the subsequent achievement of which may strike some observers as a defeat for the incumbent rather than a draw. However, given that bargaining demands are endogenous to estimates of relative power and resolve, coding such cases as victories would get at a different conceptual definition—the accuracy of belligerents’ estimates. In coding the dependent variable, I rely primarily on the Correlates of War Project’s (version 4.0) descriptions of each case (Xxxxxxx and Xxxxxx 2010). I subse- quently substantiate each coding using a number of other reference materials, includ- xxx Xxxxxxxx and Xxxxxxx’x (2005) Encyclopedia of Wars, Xxxxxxxxxx’x (2002) Warfare and Armed Conflicts, Xxxxxxx’x (2001) and Xxxxxxxxx, Xxxxxxxx, Xxxxx, Xxxxxx and Xxxxxxxx’x (2008) encyclopedias of world history, Xxxx’x (2007) Dictionary of Wars, as well as a large number of individual case histories. A codebook documenting the coding decision for each observation is included in Appendix A. As far ...
Dependent Variable. The Self-Report of Offending scale (Huizinga, Esbensen, & Xxxxxx 1991) was adapted in this dataset to measure the respondents’ reports of antisocial and illegal activities. Twenty-four items are used to assess aggressive crimes, income- generating crimes, and public order offenses. These offenses include vandalism, arson, set fire, burglary, shoplifting, received stolen property, used credit card illegally, stole car, sold marijuana, sold other drugs, carjacked, drove drunk, been paid by someone for sex, forced someone to have sex, killed someone, shot someone, shot at someone, robbery with weapon, robbery with no weapon, beaten someone, in fight, fight part of gang, and carried gun. An offending variety score was created, which represents the number of different delinquent acts committed in the previous 6 months2 -- coded from 0 (no delinquent acts) to 1.0 (all 22 acts were committed). Variety scores have been previously used to assess criminal activity. For example, Hindelang, Hirschi, and Weis (1981) use variety scores to index criminal activity and other studies have been published on the validity of variety scores (Xxxxxxxx & Xxxxxxx 1985, 1986; Xxxxxxx, XxxXxxxxx, & Xxxxxxx 2002). The variety score is used because it has the least skewed distribution of the self-report measures included in the dataset (see Appendix C).
Dependent Variable. Table 1 shows the results testing H1: Juvenile offenders who have (a) weak conventional bonds to society and (b) low self-control are more likely to be susceptible than those with strong conventional bonds to society and high self-control. Table 1 has an adjusted R2 of .09 and is statistically significant at the .000 level. Overall, the results suggest that there are several significant factors that influence susceptibility, which include parental monitoring, perceptions of chances for success, grades, self-control, gender, and ethnicity. The social bonds variable, parental monitoring, is significant holding all other variables constant. However, the relationship is in the opposite direction that is hypothesized, which suggests that juveniles with strong parental monitoring are more likely to be susceptible. I will discuss this more below. Perceptions of chances for success is also highly statistically significant. This suggests that adolescents who have a better outlook on the future are less susceptible to peer influence. In the model, grades is modestly statistically significant, holding all other variables constant. This suggests that adolescents who report better grades are less susceptible to peer influence. The other social bond variables, bonding to teachers and maternal warmth, are not significant predictors of susceptibility. The second part of the first hypothesis is also tested in this model. Self-control has a relatively strong effect when regressed onto susceptibility, holding all other constant, and the relationship is in the hypothesized direction. This suggests that juveniles with greater self-control are less susceptible to their peers. Finally, the results indicate that ethnicity, namely African American, is significant at the .01 level. The direction of the relationship suggests that African Americans are less likely to be susceptible than their white counterparts. Also, sex is statistically significant, but only at the .05 level. This suggests that males are more susceptible than their female counterparts in the sample.
Dependent Variable. The measurement of compliance is the primary challenge of this analysis. In strict legal terms, the Annex I ratifiers of the Protocol are considered to be in compliance as long as (i) they make demonstrable progress by 2005 and (ii) their emissions meet the targets averaged over the five year period of 2008-201210. It is hypothetically possible that all ratifiers might achieve their five year total commitments only in that final year, in which case, they would all be considered in compliance regardless of any over-quota emissions produced during the preceding years. However, demonstrable progress is required by 2005 and, in any event, measures taken by governments to reduce emissions take years to produce significant reductions, so it is
Dependent Variable. CONDITIONAL GRANTS PER CAPITA First, the dependent variable, the conditional grants per capita has been analysed. In 2012, a Flemish municipality received on average 173,64 conditional grants per capita. For comparison, home towns of the Flemish ministers (N=9) received on average 214,49 conditional grants per capita while the other 299 Flemish municipalities received on average 172,41 conditional grants per capita. The maximum value was 618,29 and the minimum 16,68. In 2013, a Flemish municipality received an average 176,42 conditional grants per capita while home towns of Flemish ministers received an average of 222,33 and the remaining 299 municipalities 175,03 conditional grants per capita. The municipality that received the most conditional grants per capita got 614 while the minimum was 16,56. The results thus show that on average, the home towns of Flemish ministers do receive more conditional grants per capita. However, in the nine home towns, a larger standard deviation was also noticed. A summary of the dependent variable is given in Table 2. 2012 Mean Std. Dev. Min. Max. All Flemish municipalities 173,64 100,09 16,68 618,29 Home towns of Flemish ministers 214,49 160,21 51,3 519,30 299 other municipalities 172,41 97,88 16,68 618,29 2013 All Flemish municipalities 176,42 106,67 16,56 614 Home towns of Flemish ministers 222,32 175,19 30,68 532,17 299 other municipalities 175,04 104,08 16,56 614 Table 2 Descriptive results section 4.1.2
Dependent Variable. Political Order My dependent variable will measure the concept of political order, which, as mentioned earlier, measures three aspects of personal security: for one’s life, family, and source of livelihood (North, Summerhill, Xxxxxxxx). Using theory from economics and 6 PRI= Partido Revolucionario Institucional PAN= Partido Accion Nacional PAS= (former) Partido Alianza Social PT= Partido del Trabajo PRD= Partido de la Revolucion Democracatica PVEM= Partido Verde Ecologista de Mexico CONV= Convergencia Other= any party or alliance party not included above 7 In particular cases where there was a candidate from an alliance party(ie: PRI/Alianza), I only coded that party as a separate party if there was also a candidate from the non-alliance form of the party (ie: PRI). sociology, North, Summerhill, and Xxxxxxxx argue that political order exists when citizens find it in their best interest to obey (and in some cases, enforce) the formal rules in society (4). To measure political order, I chose my dependent variable to be the number of annual homicides per 100,000 people in each municipality between 2001 and 2010. Homicide is a measure of overall crime, and therefore an increase in homicide represents a decrease in political order. The municipality’s crime rate is a valid indicator of political order, because given that the government’s first and foremost obligation to its citizens is to provide security, crime represents the government’s failure to meet its obligations and provide order. Although actual crime rates do not fit the definition of the political order given by North, Summerhill, and Xxxxxxxx, I believe it is an appropriate measure of the political order in a given region for the above reasons. While there are data measuring the perception of crime in Mexico, the sample is not large enough to disaggregate at the municipal level. I obtained my homicide data from a spreadsheet that was made by a blogger, Xxxxx Xxxxx, whose self-proclaimed interests are data analysis and information systems. Although not ideal, Xxxxx obtains his homicide data from the Mexican National Institute of Statistics, Geography, and Informatics (INEGI) between 1990 and 2010 and his population data from the Mexican census.8 INEGI compiles these homicide statistics based on death certificates which, unlike estimates provided by law enforcement agencies, are not affected by state level discrepancies in the legal definition of homicide or prosecutors’ biases, and decrease the issue...
Dependent Variable. Language Regime Choice The dependent variable is trichotomous. Specifically, it identifies whether the educational language regime of country j in year k is power-concentrating, power-sharing, or power-neutralizing. The variable Language Regime Choice is coded as power-concentrating (assigned a value of 0) if the education system is monolingual in the dominant group language; power-sharing (assigned a value of 1) if multilingual; and power-neutralizing (assigned a value of 2) if monolingual in a lingua franca. As with the dependent variable in the previous section, the data is from a mix of sources, xxxx Xxxxxxx (2007) providing the initial information for most entries. Since language regimes are political institutions, a lag is to be expected between the political shock and the actual choice of language regime. Some lags are short. In Brunei, for example, the National Program of Education was passed within one year of independence 26 For these latter cases, I consider this second round of political independence as a new observation. Anecdotal evidence suggests that even within the same country, dominant groups behave differently during the two periods given a different likelihood of remaining in power (see for exaxxxx, Raun 2001).
Dependent Variable. The dependent variables were measured using a 113-item questionnaire, measuring a total of 6 constructs. The questionnaire was divided into three consecutive blocks: appropriateness-rating of the robot-group scenes, cultural- and personality background questions, and general demographic questions. In the first block, participants were asked to rate the 37 'robot approaches a family' scenes which have been described in the previous section. To avoid order-effects, all scenes were randomized. Participants were provided with the instruction: “The robot approached the family and has come to a halt between particular family members at a particular distance. Now it will interact with them”, and asked to indicate on a 7-point Likert scale whether or not 1 xxxx://xxxxx.xx//ergonomics/ Figure 4: F-Formation used. Dark grey indicates possible location of the robot. Grey: intimate zone, light grey: personal space zone. Participants standing in a circle with a diameter of 122 cm. Under review the position of the robot was considered appropriate. Another four items were included in this block to measure how participants themselves would approach the family. Two items were included to check the approach position- and distance manipulation. A final item was included in which we asked participants if they could indicate where they thought the family they had seen in the situations originated from. The second block of the questionnaire consisted of a series of validated scales measuring four dependent variables. Participants’ general attitude towards robot was measured by the Negative Attitude Towards Robots scale, a 14-item 7- point Likert scale. One way to explain cultural differences is by measuring individual vs. group self-representations. This was operationalized using 7 items, by Xxxxxx & Xxxx [1], and analyzed in a similar way as Xxxx et al. [22]. An indication of whether participants were members of a high- contact, or low-contact culture was assessed by measuring closeness. Five items from the IPROX (iconic proximity) questionnaire were used [7]. The final construct in this block was personality. We measured the Big Five personality traits using the 20-item Mini-IPIP scale [2]. The final block of questions included demographic questions like gender, age, nationality, and level of education. Social-demographic questions like nationality of ancestors, marital status and children were also included. Participants Participants were recruited from three different count...
Dependent Variable. The starting point of my dependent variable measurement is the definition of contagion that I have assumed. To recap, Kodres and Xxxxxxxx (2002, 772) define 13 See Appendix III for the distribution of daily data by country and SDDS subscription status; see Appendix IV for descriptive information on each nation in the sample, including SDDS subscription details, geographical region and market type. $" contagion as “a general price movement in one market resulting from a shock in another market.” Given the authors’ specification of price movement, my measurement of contagion must be based on the value of a financial instrument. I have chosen government bonds as the study’s financial stability indicator since bond returns are determined by investors’ perception of investment risk. Government bonds have been issued and their returns quoted for a significant amount of time, so bond data is readily available. Furthermore, bonds are essential to government operations and are issued and traded regularly. Accordingly, the bond market is highly liquid and bond returns are generally accurate, especially in emerging economies where trading volumes are high. Alternative measures of country risk that were considered were stock market indices and credit default swap (CDS) rates. These instruments would be legitimate measures of volatility, but this paper only uses bond data to keep the scope tractable. The metric I have chosen to measure sovereign bond returns is XX Xxxxxx’x Emerging Market Bond Index (EMBI). The index is a bond return index, so lower values denote increased risk. The EMBI’s “[w]ell-defined liquidity criteria ensure the index provides a fair and replicable benchmark,” so there is no need to control for market illiquidity or trading volumes (XX Xxxxxx). Finally, the EMBI has been widely employed in past research on contagion (Xxxxxxxx and Schmukler 2002; Alexander, et al. 2008; Glennester and Shin 2003; Xxxxxxxx, et al. 2003). A restriction of using the EMBI is that my sample is limited to a group of emerging markets that issue bonds and are quoted by the EMBI. However, these emerging markets are of the most interest in my theory and exhibit cross-country, cross-time variation on SDDS membership. $# To arrive at the final analysis of my contagion variable there were a series of intermediate steps. First, I transformed the EMBI index into daily percent change values. Then I conducted a regression of these daily changes with a variety of explanatory variabl...
