independent sample t-test classical design used in psychology/medicine n subjects are randomly...
Post on 21-Dec-2015
218 views
TRANSCRIPT
Independent Sample T-test
• Classical design used in psychology/medicine • N subjects are randomly assigned to two
groups (Control * Treatment). • After treatment, the individuals are measured
on the dependent variable. • A test of differences in means between groups
provides evidence for the treatment's effect.
Measures of Variation
• A lot of statistical techniques (using interval data) use measures of variation in some manner
• What is the difference between a standard deviation, the standard error of the mean, and the standard error of the difference between means? Or How are they related? Look in the glossary to help you answer these questions?
Using Measures of Variation
• Leaned how to measure variation in data, i.e., variance, standard deviation (Ch.4)
• Used the normal curve & SD to calculate z-scores and probabilities (Ch.5)
• Used the normal curve & the z-score & the SE of the mean to calculate confidence intervals (Ch.6)
• Used the concept of the confidence interval and the standard error of the differences between means to calculate the t-test (Ch.7)
• Use the sum of squares Σ(X – Mean)2 [sum of the squared differences from the mean] in ANOVA
Null Hypothesis
• The two groups come from the same population or that the two means are equal
• μ1 = μ2
Levels of Significance
• What does an α = .05 level of significance mean?
• We decide to reject the null if the probability is very small (5% or less) that the sample difference is a product of sampling error.
• The observed difference is outside the 95% confidence interval of the difference
Choosing a Level of Significance
• Convention• Minimize type I error – Reject null
hypothesis when the null is true• Minimize type II error – fail to reject null
when the null is false• Making alpha smaller reduces the
likelihood of making a type I error• Making alpha larger reduces the
probability of a type II error
Independent Sample T-test Formula
21
21
21
222
211
221 NN
NN
NN
sNsNs xx
t =
21
21
XXs
XX
Assumptions of the t-test
• 1. All observations must be independent of each other (random sample should do this)
• 2. The dependent variable must be measured on an interval or ratio scale
• 3. The dependent variable must be normally distributed in the population (for each group being compared). (NORMALITY ASSUMPTION) [this usually occurs when N is large and randomly selected]
• 4. The distribution of the dependent variable for one of the groups being compared must have the same variance as the distribution for the other group being compared. (HOMOGENEITY OF VARIANCE ASSUMPTION)
Don’t worry about these assumptions to much, but
• Point 1: statistical tools are attempting to quantify and analyze very complex social/political phenomenon
• Point 2: For these test to be accurate they relay on simplifying the world with many assumptions that might not be true
• Point 3: social science researchers violate these assumptions quite often, but try to be honest about it
• Point 4: there are sometimes ways of testing and adjusting for violations
SPSS & the Independent Sample T-Test
Independent Samples Test
.073 .789 .288 18 .777 4.00000 13.91242 -25.22892 33.22892
.288 17.971 .777 4.00000 13.91242 -25.23230 33.23230
Equal variancesassumed
Equal variancesnot assumed
VAR00001F Sig.
Levene's Test forEquality of Variances
t df Sig. (2-tailed)Mean
DifferenceStd. ErrorDifference Lower Upper
95% ConfidenceInterval of the
Difference
t-test for Equality of Means
Group Statistics
10 102.0000 31.72801 10.03328
10 98.0000 30.47768 9.63789
VAR000021.00
2.00
VAR00001N Mean Std. Deviation
Std. ErrorMean
Crime and Support for Democracy
Kenneth E. FernandezDepartment of Political Science
University of Nevada, Las Vegas
Michele KuenziDepartment of Political Science
University of Nevada, Las Vegas
Research Question
Does crime influence how citizens evaluate democracies or influence their
level of support for democracy?
Why is this an Interesting Question?
• Scholars have noted that few studies, especially in the field of political science, have examined crime and its impact on society in a comparative manner (Beirne 1997; Howard, et al. 2000).
• Current public opinion research suggests that the issue of crime is of great concern to citizens across the globe (Quann and Hung 2002)
• Many countries that have experienced recent democratic reforms have also experienced increasing crime rates (Bergman 2006).
What theories inform the study of crime and democracy?
–Social Capital
–Rational Choice
–Modernization Theory
Social Capital Theories: – Theory suggests that attitudes and feelings of
the masses regarding obligations within a group, civic attachments, trust, and efficacy are important determinants of democratic transition and sustainability (Muller and Seligson 1994).
– Yet crime and violence is likely to reduce trust and social capital (Lederman et al 2002).
Rational Choice: – Citizens are treated as utility maximizing
consumers– A political regime is expected to produce some
value or benefit to the consumer (Lane 1995).– Regime support depends on how well citizens’
demands are met– Thomas Hobbes argues that protection against
violence is the highest and most fundamental human need and a citizen of the state has a right to revoke its support and loyalty when the state proves incapable of protecting its subjects.
The Paradox of Modernization theory: – Modernization leads to increasing wealth,
leading to more literacy, education, urbanization, leading to democratization because citizens are less willing to put up with oppressive, authoritarian behavior and more willing to support democracy (Lipset 1959)
– But modernization may cause a breakdown in traditional structures and values, an increase in interaction, tension, and conflict. These social changes then contribute to rising criminal activity (Heiland and Shelley 1992).
Modernization Theory, Crime & Democracy– Consolidating both the political science and the
criminology/sociology literature on modernization theory, we argue that socio-economic conditions, as well as, crime and public safety influence citizens’ support for democracy.
Crime is Often Ignored as a Factor in Democratization
– Much of the prior research on the development of democracy focuses on economic conditions (Arat 1988; Epstein et al. 2006; Lipset 1959; Przeworski et al. 2000).
– In fact, Kugler and Feng describe the recent application of modernization theory as positing that “economic development is a sufficient, rather than a necessary, condition for democratic transitions” (1999, 140).
– We argue that crime and public disorder may pose even greater challenges to the legitimacy and effectiveness of democratic institutions
Methodology: Macro or Micro Level Analysis?– Much of the research testing the modernization
theory uses a macrolevel approach utilizing the nation-state at time t as the unit of analysis. Aggregate national characteristics (i.e., per capita GDP) are then used as predictors of some measure of democracy.
– Yet Lipset (1959) posits that increases in wealth and income leads to democratic transition because it affects citizens’ values.
– We would like to examine this microlevel statement.
Microlevel Analysis:
– An examination of citizen attitudes is appropriate because the success of democratic consolidation in these emerging democracies is often linked to citizens’ support for democracy and perceptions of government performance (Hiskey and Bowler 2005; Lagos 2001; Diamond 1999; Lipset 1959).
DATA: • Afrobarometer (AB) and Latinbarometro (LB)
Survey Data• The AB and LB together contain 33 countries (16
in Africa and 17 in Latin America).
Two Statistical Approaches– Analysis of all countries pooled together (over
18,000 respondents in African and 13,000 in Latin America
– Analysis of 4 cases
Case Selection:– Initial Case - NIGERIA: Like many other
African countries, Nigeria experienced a transition to democracy in the 1990s. We chose Nigeria because crime is a very salient issue in this country, and Nigeria has been well known for mob justice or what has been called “jungle justice”
– Most Different Systems Design: We performed a cluster analysis to find a case in the Afrobarometer data that was “most different” to Nigeria in relation to three factors: crime, democracy and per capita GDP.
Cluster Analysis:
The Euclidean distance between two cases is calculated by finding the square root of the sum of the squared differences between the two cases for each of the three variables (crime rate; per capita GDP; level of democracy):
222jijiji zzyyxx
MALAWI: Nigeria has a per capita GDP nearly 3 times that of Malawi, but Nigeria has a substantially higher reported victimization rate (Nigeria had the highest reported attacks of the 14 African cases and Malawi had the lowest). Both countries were given a score of 4 for political rights and civil liberties by Freedom House in 2005.
CHILE & NICARAGUA: Both Chile and Nicaragua have modest crime rates compared to the rest of Central and South America, but differ greatly on GDP and level of democracy. Chile has a per capita GDP almost 7 times higher than Nicaragua and Chile has a score of 1 for both political rights and civil liberties while Nicaragua scored a 3 for both.
THE 4 CASES: The results seem to make intuitive sense. There is substantial variation in GDP across cases, they vary in crime rates from low to moderate to high levels of victimization, and levels of political freedom and civil liberties varied from the highest (score of 1) to moderate/low (score of 4).
Operationalization
Dependent Variables: Attitudes toward Democracy Support for Democracy • Which of these three statements is closest to your
own opinion? • A: Democracy is preferable to any other kind of
government. • B: In some circumstances, a non-democratic
government can be preferable. • C: For someone like me, it doesn’t matter what kind
of government we have. • Recoded so that the value on this variable is 1 if the
response was A, and 0 otherwise.
• Satisfaction with Democracy • Overall, how satisfied are you with the way
democracy works in __________.? Are you:• 0=Tanzania/Nigeria is not a democracy,
1=Not at all satisfied, 2=Not very satisfied, 3=Fairly satisfied, 4=Very satisfied.
Chile and Nicaragua:• In general, would you say that you are very
satisfied, fairly satisfied, not very satisfied or not at all satisfied with the way democracy works in (country)? 4=Very satisfied - 1=Not at all satisfied
Independent Variables
• Experience with Crime
• Comparative assessments of safety from crime and violence
• Crime Performance
Control Variables
• Demographic variables: gender, level of poverty, age, education
• Other variables: political engagement, identification with the party in power, and subjective assessments of satisfaction with life and the economy and government service delivery
TABLE 3: Attitudes Toward Democracy
AFRICA
Support Democracy Satisfaction with Democracy Nigeria 2000 81% 85% Nigeria 2001 73 57 Nigeria 2003 69 34 Malawi 1999 65 57 Malawi 2003 66 48 LATIN AMERICA
Chile 1996 54% 29% Chile 2000 55 34 Chile 2003 52 34 Chile 2004 62 44 Nicaragua 1996 64 25 Nicaragua 2000 68 17 Nicaragua 2003 59 34 Nicaragua 2004 50 23
Table 1: Pooled Afrobarometer and Latinbarometro Data Support for
Democracy (AB Pooled Data)
Satisfaction w/ Democracy
(AB Pooled Data)
Support for Democracy
(LB Pooled Data)
Satisfaction w/ Democracy
(LB Pooled Data) Est. Method
Logit
Ordered Logit
Logit
Ordered Logit
Feel Safer
0.200*** (0.036)
0.267*** (0.038)
0.146* (0.084)
0.325*** (0.076)
Victim of Crime
-0.284*** (0.061)
-0.145*** (0.043)
0.023 (0.056)
-0.121** (0.059)
Discuss Politics
0.127*** (0.024)
0.007 (0.020)
0.199*** (0.053)
0.051 (0.032)
Radio-News
0.078*** (0.028)
0.050** (0.026)
0.033 (0.020)
0.033* (0.013)
Service Delivery
0.180*** (0.030)
0.374*** (0.036)
0.030 (0.030)
0.163*** (0.030)
Economy
0.037 (0.033)
0.251*** (0.061)
-0.016 (0.050)
0.366*** (0.055)
Quality of Life
-0.041 (0.040)
0.059 (0.041)
0.106 (0.068)
0.229*** (0.054)
Education
0.023 (0.022)
-0.090*** (0.022)
0.035** (0.014)
-0.001 (0.009)
Poor
-0.010 (0.062)
-0.087** (0.045)
0.032 (0.053)
0.041 (0.063)
Female
-0.118*** (0.048)
-0.047 (0.049)
-0.067 (0.050)
-0.084 (0.049)
Age
0.004* (0.002)
0.005*** (0.002)
0.011*** (0.003)
0.004 (0.002)
Cons
-0.828*** (0.315)
-1.751*** (0.407)
# obs
18,476
18,695
13,826 14,125
Pseudo R2 0.034 0.061 .0214 .0435 LR chi2 235.10 904.72 153.90 335.45 Prob > chi2 0.000 0.000 0.000 0.0000
TABLE 4: ATTITUDES TOWARD DEMOCRACY Model 1 Model 2 Support for
Democracy (Nigeria)
Support for Democracy (Malawi)
Satisfaction w/ Democracy (Nigeria)
Satisfaction w/ Democracy (Malawi)
Est. Method
Logit Logit Ordered Logit Ordered Logit
Feel Safer
0.236*** (0.044)
-0.009 (0.051)
0.385*** (0.038)
0.078* (0.045)
Victim of Crime
-0.499*** (0.113)
-0.627** (0.261)
-0.069 (0.101)
0.182 (0.228)
Discuss Politics
0.058 (0.039)
0.054 (0.049)
0.060* (0.033)
-0.060 (0.041)
Radio-News
0.144*** (0.046)
-0.056 (0.051)
0.002 (0.041)
0.163*** (0.044)
Service Delivery
0.101** (0.044)
0.166*** (0.048)
0.179*** (0.038)
0.488*** (0.044)
Economy
0.081* (0.044)
0.085 (0.070)
0.428*** (0.038)
0.240*** (0.059)
Quality of Life
-0.112*** (0.046)
-0.005 (0.073)
0.173*** (0.040)
0.014 (0.062)
Education
-0.018 (0.025)
0.115** (0.051)
-0.049** (0.021)
-0.173*** (0.041)
Poor
-0.037 (0.061)
-0.125 (0.108)
-0.157*** (0.051)
-.343*** (0.091)
Female
-0.052 (0.099)
-0.412*** (0.144)
-0.019 (0.084)
-0.449*** (0.122)
Age
0.002 (0.004)
0.005 (0.005)
-0.000 (0.003)
0.008* (0.004)
Cons
-0.440 (0.326)
0.146
(0.494)
# obs
2,261
1,009
2,270
1,019
Pseudo R2 0.036 0.032 0.104 0.088 LR chi2 99.63 41.26 627.93 258.38 Prob > chi2 0.000 0.000 0.000 0.000
TABLE 5: ATTITUDES TOWARD DEMOCRACY II Model 3 Model 4 Support for
Democracy (Nigeria)
Support for Democracy (Malawi)
Satisfaction w/ Democracy (Nigeria)
Satisfaction w/ Democracy (Malawi)
Est. Method
Logit
Logit
Ordered Logit
Ordered Logit
Govt. Reducing Crime
0.197*** (0.056)
0.18 (0.78)
0.531*** (0.048)
0.324*** (0.069)
Victim of Crime
-0.522*** (0.112)
-0.658*** (0.258)
-0.114 (0.101)
0.248 (0.226)
Discuss Politics
0.064* (0.038)
0.060 (0.049)
0.056* (0.033)
-0.053 (0.042)
Radio-News
0.144*** (0.046)
-0.057 (0.051)
0.008 (0.040)
0.156*** (0.045)
Service Delivery
0.127*** (0.043)
0.157*** (0.048)
0.209*** (0.037)
0.481*** (0.044)
Economy
0.089** (0.043)
0.094 (0.070)
0.421*** (0.038)
0.240*** (0.060)
Quality of Life
-0.120*** (0.046)
-0.021 (0.073)
0.145*** (0.040)
0.004 (0.062)
Education
-0.010 (0.025)
0.114** (0.051)
-0.045** (0.021)
-0.153*** (0.042)
Poor
-0.064 (0.060)
-0.144 (0.109)
-0.215*** (0.052)
-0.358*** (0.091)
Female
-0.047 (0.098)
-0.417*** (0.144)
-0.048 (0.083)
-0.470*** (0.123)
Age
0.003 (0.004)
0.006 (0.005)
0.002 (0.003)
0.009** (0.005)
Cons
-0.270 (0.323)
0.143
(0.505)
# obs 2,266 1,008 2,278 1,014 Pseudo R2 0.029 0.033 0.109 0.094 LR chi2 82.51 42.29 660.51 273.63 Prob > chi2 0.000 0.00 0.000 0.000
TABLE 6: ATTITUDES TOWARD DEMOCRACY I Model 5 Model 6 Support
Democracy (Chile)
Support Democracy (Nicaragua)
Satisfaction w/ Democracy
(Chile)
Satisfaction w/ Democracy (Nicaragua)
Est. Method
Logit
Logit
Ordered Logit
Ordered Logit
Feel Safer
0.6105*** (0.1449)
0.3072** (0.1441)
0.7250*** (0.1254)
0.5156*** (0.1358)
Victim of Crime
0.0789 (0.1843)
-0.3046 (0.1879)
-0.0113 (0.1725)
0.1028 (0.1651)
Discuss Politics
0.3767*** (0.1111)
0.3704*** (0.0987)
0.0182 (0.1255)
0.0377 (0.0863)
TV-News
-0.047
(0.0545) -0.04405
(0.066903) 0.0299
(0.0522) -0.0222 (0.0570)
Service Delivery
-0.015
(0.0617)
-0.123** (0.0535)
0.1971** (0.0648)
0.1420*** (0.0524)
Economy
0.3088** (0.1231)
-0.0029 (0.1174)
0.6488*** (0.1303)
0.3951*** (0.0940)
Quality of Life
0.2837** (0.1201)
0.0178 (0.1067)
0.4262*** (0.1164)
0.2818*** (0.1007)
Education
0.0387 (0.0295)
-0.0118 (0.0228)
0.0491* (0.0294)
0.0045 (0.0193)
Poverty
0.1861 (0.1395)
-0.0366 (0.1061)
0.1574 (0.1480)
0.004932 (0.0889)
Female
0.0174 (0.180683)
-0.2390 (0.1811)
-0.089 (0.1734)
-0.65567*** (0.1640)
Age
0.0111* (0.0063)
0.0120* (0.0064)
0.0180** (0.0061)
0.011286* (0.0061)
Cons
-4.5247*** (0.7308)
-0.6752 (0.6975)
# obs 1023 613 1040 652 Pseudo R2 0.1020 0.0478 0.1292 0.0585 LR chi2 85.54 30.44 165.28 81.41 Prob > chi2 0.0000 0.0014 0.0000 0.0000
TABLE 7: ATTITUDES TOWARD DEMOCRACY II Model 7 Model 8 Support
Democracy (Chile)
Support Democracy (Nicaragua)
Satisfaction with Democracy
(Chile)
Satisfaction with Democracy (Nicaragua)
Est. Method
Logit
Logit
Ordered Logit
Ordered Logit
War on Crime
0.3019** (0.1336)
0.3320*** (0.1180)
0.4495*** (0.1365)
0.3620*** (0.1107)
Victim of Crime
-0.0176 (0.1896)
-0.2076 (0.1929)
-0.0202 (0.1776)
0.1336 (0.1672)
Discuss Politics
0.4759*** (0.1216)
0.3663*** (0.1002)
0.0278 (0.1278)
0.0183 (0.0856)
TV-News
-0.0327 (0.0568)
-0.0232 (0.0679)
0.0366 (0.0507)
0.0227 (0.0581)
Service Delivery
0.0182
(0.0648)
-0.1179** (0.0561)
0.2107*** (0.0662)
0.1414*** (0.0526)
Economy
0.4059*** (0.1296)
0.0094 (0.1179)
0.7059*** (0.1353)
0.4070*** (0.0937)
Quality of Life
0.2248* (0.1226)
-0.0108 (0.1104)
0.4065*** (0.1146)
0.2658*** (0.1028)
Education
0.0384 (0.0309)
-0.0092 (0.0234)
0.0484 (0.0305)
0.0066 (0.0195)
Poverty
0.1871 (0.1437)
-0.0438 (0.1099)
0.2208 (0.1464)
0.0152 (0.0908)
Female
0.0260 (0.1852)
-0.2137 (0.1839)
-0.0736 (0.1750)
-0.6451 (0.1668)
Age
0.0080 (0.0062)
0.0143** (0.0065)
0.0149*** (0.0057)
0.0114* (0.0064)
Cons
-4.1139*** (0.7574)
-0.9637 (0.6937)
# obs 1000 604 1018 640 Pseudo R2 0.0903 0.0538 0.1159 0.0548 LR chi2 74.65 32.55 133.45 73.55 Prob > chi2 0.0000 0.0006 0.0000 0.0000
Results
Attitudes toward Democracy
Feeling Safe
• Nigerians who report feeling better off in terms of safety from crime and violence are significantly more likely to:
– Support democracy
– Be satisfied with the way democracy is working
• Malawians who report feeling better off in terms of safety from crime and violence are significantly more likely to:
– Be satisfied with the way democracy is working
• Nicaraguans and Chileans who felt the country was getting safer were also more likely to:
– Support democracy
– Be satisfied with the way democracy is working
Government Performance in Crime Reduction
• Results generally stronger
Crime Victimization
• Nigerians and Malawians who report a higher number of experiences with crime victimization have significantly lower levels of support for democracy.
• There appears to be no relationship between crime victimization and attitudes toward democracy in the Chile and Nicaragua countries.
Conclusion and Implications
• Perceptions of safety and government performance in the area of crime affect attitudes toward democracy.
• Crime victimization generally does not appear related to satisfaction with democracy.
• Issue of crime likely to affect the consolidation of democracy
• Future research should:
–Examine other regions
–Assess whether there is a threshold effect
–Examine potential intervening variables