Midterm Flashcards
Hypothesis
Answer to the research question that may or may not be right
Testable Hypothesis
A version of your hypothesis that you can test with empirical data and evidence
Causal Mechanism
Showing a relationship between two variables. How they are related
Independent Variable
X; phenomenon doing the explaining
Dependent Variable
Y; phenomenon being explained
Internal Criteria
How does the hypothesis fare?
Falsifiability
The hypothesis must be capable of being proven wrong
Parsimony
Hypothesis must be able to explain while using as few explanatory variables as possible
Encompassing
Hypothesis must be able to apply to a lot of different cases
Concrete
Concepts cannot be too abstract, must be clear and have sound reasoning
Operational Definition
Very specific definitions of variables that will enable us to gather data
Reliability
When testing the hypothesis, the same result must be obtained if the measurement process is repeated
Validity
The variables accurately reflects the abstract concept of interest.
Validity implies reliability, but NOT the other way around.
Scales of Measurement
AKA Measure Precision
Nominal
With qualitative data, unordered categories
ie; single, married, divorced
Ordinal
With quantitative or qualitative data, ordered categories
ie; social class, type of education
Interval
With quantitative data, numeric values with a significant distance between them
ie; amount of education, average temp
Qualitative Data
Numbers are not involved
Quantitative Data
Numbers are involved, central to data
Categorical
Has categories to classify data
ie; type of education, marital status
Non-categorical
ONLY numbers, with interval data
Discrete Data
Data with a definite number or count
Continuous Data
An infinite number of values can be measured
ie; ages
Central Tendency
Identifying frequently occurring values
ie; mean, median, and mode
Mean
Add up values and divide by number of cases
Use the mean when you have data that does not include extreme scores and are not categorical
ie; “Average”
Median
The midpoint in a set of scores
value of the ((n+1)/2)th case (n is # of cases)
Most commonly used for ordinal scale variables
Use the median when you have extreme scores and you don’t want to distort the average
Mode
Most frequently reported/utilized category in a data set
Us the mode when the data is categorical
Dispersion
Reflects how data points differ from one another
ie; range, variance, standard deviation
Range
Difference between the values of the smallest and largest variables
Variance
Just know how to interpret
Bigger values mean more dispersion, but otherwise its kind of hard to interpret
Standard Deviation
Average difference/distance from the mean
Bigger numbers mean more dispersion
Scatterplot
For ordinal and interval scale variables
“Chart Junk”
Un-necessary chart material that distracts from the data and makes it difficult to interpret
Graphical Integrity
Bad graphs misrepresent data through distortion, changing scale of measurement, etc. which lacks integrity
Pearson’s Correlation Coefficient
Quantifies strength and direction of relationship of variables
Ranges from -1 to 1
The larger the ABSOLUTE VALUE, the stronger the relationship
Positive sign indicated a positive relationship and vice versa
Coefficient of Determination
R^2 - Proportion of shared variance
Ranges from 0 (0%) to 1 (100%)
Square of Pearson’s Correlation Coefficient
Spearman’s Correlation Coefficient
Similar to Pearson’s in all ways
EXCEPT—can be used for either two ordinal scale variables or for one ordinal and one interval scale variable
Regression Line/Linear Regression
Trend lines in a graph
Works for any combination of ordinal and interval scale variables
Y = a + bx + e
Linear Regression Model
y
dependent variable
x
independent variable
a
y-intercept of the line
b
slope of the line
slope
rise/run
e
random error (not always there)
Ordinary Least Squares
Estimation of a Regression Line
The line with the smallest squared distance
Regression Prediction Equation
Y’ = a + bx
Slope Coefficient
the “b” in “Y’ = a + bx”
Predicted change in Y as X increases by 1 unit
Crosstab
For two qualitative variables
conditional distributions identical = no association
conditional distributions differ = association
Chi-Squared Statistic
For two qualitative variables
minimum value = 0 : NO association
Non-zero values indicate an association: a bigger number equals a stronger association