research design quiz #12 Flashcards
After we have gathered and cleaned our data, we can run different
statistical analyses to…
- Explore relationships between different variables
- Understand the distributions of our variables
- Test measures of central tendency and dispersion
- Answer research questions or hypotheses
Different types of statistics are used to answer different types of
questions:
1)M of a
2) C s
3) I s
- Measures of association
- Comparative statistics
- Inferential statistics
Measures of _________ determine the strengths of _________ among the variables that are being studied
association;
relationships
two continuous measures (measures from -1 = perfect negative association to +1 = perfect positive association)
which of the seven (7) measures of association is this one?
Pearson’s R
measures the association between two dichotomies
which of the seven (7) measures of association is this one?
Phi coefficient
measures the association between two rank ordered
variables
which of the seven (7) measures of association is this one?
Spearman’s rank order correlation (Rho)
measures the association between a continuous variable and a
dichotomous variable
which of the seven (7) measures of association is this one?
Point-biserial correlation
measures the association between a continuous variable and a theoretically continuous-but-polytomized variable categorized into 3+ levels (think Likert scale scores)
which of the seven (7) measures of association is this one?
Polyserial correlation
measures the association between two dichotomous variables and
estimates what the association would be had both variables been continuous
which of the seven (7) measures of association is this one?
Tetrachoric correlation
measures the association between ordinal variables with two or
more levels
which of the seven (7) measures of association is this one?
Polychoric correlation coefficient
_______ ________ involves analysis of the attributes of
the variables being examined to better describe the relationship
between two variables
comparative statistics
First comparative statistic:
Frequently presented in research
Usually reflects the number of crimes divided by the population multiplied by 100,000
This equation gives you the crime count per 100,000 people; could be per million, ten thousand, thousand, hundred etc
crime rates
Second comparative statistic:
Same as above, except that the chosen denominator is not the population count, but rather a factor related to the crime type
For example, dividing the number of burglaries by the number of households; number of MVTs by number of registered vehicles in an area and so on
crime-specific rates
Third comparative statistic:
Great for comparing data fluctuations over time
Easy to calculate: Subtract the first-time value from the second-time value, then divide by the first-time value
Percentage change
fourth comparative statistic:
Great for examining the behavior of data over time
Great for comparing similar and/or different data side by side over time
Can apply lines of best fit (linear, average, polynomial etc.)
Good for assessing the impact of crime prevention strategies
Trend analysis
_____ _______ involves not only describing relationships,
but makes predictions and/or inferences about dependent variables based on the influence of independent variables
inferential statistics
We have many different types of inferential statistics to choose from:
1) B a
2) C t
3) t-t
4) C
5) A
6) B r
7) M r
- Bivariate analyses
- Contingency tables
- t-tests
- Correlations
- ANOVAs
- Bivariate regression
- Multivariate regression
INFERENTIAL STATS:
Which for the seven (7) inferential stats should we use for this?
Examining the relationship between to variables
b a
Bivariate analyses
INFERENTIAL Stats:
Which for the seven (7) inferential stats should we use for this?
Usually used when you have two nominal variables you want to compare
Can simply use cross-tabulation tables
Can also use Chi-square (2) to test for statistical significance
c t
Contingency tables
INFERENTIAL Stats:
Which for the seven (7) inferential stats should we use for this?
Used to test whether group means are statistically significantly different from each other
Independent samples ____ versus paired(aka dependent) samples _____
t-t
t-tests
INFERENTIAL Stats:
Which for the seven (7) inferential stats should we use for this?
Again, used for assessing the relationship between two variables (review previous slide on correlation types)
If we square the correlational value, we have the proportion of explained variance
correlations
INFERENTIAL Stats:
Which for the seven (7) inferential stats should we use for this?
Principally used to determine whether three or more groups significantly vary in
relation to each other
We can determine if the mean values of a measure for each group are
significantly different from each other
Example: we have three groups of officers and want to know whether the
number of arrests significantly differ between groups
A
ANOVA’s (analysis of variance)
what does ANOVA stand for
analyisis of variance
INFERENTIAL Stats:
Which for the seven (7) inferential stats should we use for this?
Based on the principle that over time things tend to regress toward the mean
Use one variable (IV) to predict another variable (DV)
Several assumptions must be met in order to use linear regression (Normality, linearity, homoscedasticity)
B r
Bivariate regression
INFERENTIAL Stats:
Which for the seven (7) inferential stats should we use for this?
Same as above, but you are assessing the predictive potential of multiple IVs and control variables in relation to a dependent variable
m r
multivariate regression