SOWO 911 Flashcards
1
Q
What is a research question?
A
- A question that defines what a study hopes to learn
- Narrow and specific
- Can be answered using observable evidence (data)
- Significant/relevant to policy, theory, or practice
- Not a statement
2
Q
What is a directional hypothesis?
A
A directional (or one tailed hypothesis) states which way you think the results are going to go, an example of a directional hypothesis is:
Adolescents who have higher levels of social support system are more likely to take an HIV test than adolescents who have lower levels of social support system.
- A directional hypothesis is one-tailed.
- It is used if the theory and empirical evidence implies direction.
3
Q
What is a non-directional hypothesis?
A
- A non-directional hypothesis is two-tailed.
- It is used if the theory and empirical evidence does not imply a direction, only a relationship.
- Critical region split on both sides of the sampling distribution ( 2.5% on each side)
4
Q
What is a research hypothesis?
A
A research hypothesis is a specific, clear, and testable proposition or predictive statement about the possible outcome of a scientific research study based on a particular property of a population, such as presumed differences between groups on a particular variable or relationships between variables.
- Testable form
- Three types of research hypothesis (Ranjit, 2011 Research Methodology):
- Hypothesis of difference: A hypothesis in which a researcher stipulates that there will be a difference but does not specify its magnitude.
- Hypothesis of point-prevalence: When a researcher has enough knowledge about a phenomenon that he/she is studying and is confident about speculating almost the exact prevalence of the situation or the outcome in quantitative units.
- Hypothesis of association: When a researcher have sufficient knowledge about a situation or phenomenon and are in a position to stipulate the extent of the relationship between two variables and formulate a hunch that reflects the magnitude of the relationship.
5
Q
What is a statistical hypothesis?
A
- All research hypotheses can be written using statistical terms, where:
- H_0 is the null hypothesis (which states that the research hypothesis is not supported)
- H_1 is the alternative hypothesis (which states that the research hypothesis is supported)
- H_0 and H_1 are mutually exclusive
- Must include all possible values of the parameter
6
Q
Rejection regions
A
- Values on the sampling distribution of the test statistic that have a probability equal to or less than alpha that the null hypothesis is true. If the test statistic falls into the rejection region, the null hypothesis is rejected.
- The rejection region is the area under the curve that falls above (or below, for the left side) the critical value
- It represents the probability of an observed result arising by chance
7
Q
What does a non-directional hypothesis imply about the rejection region?
A
- Non-directional hypotheses have two-tailed rejection regions
- For two-tailed rejection regions, you split the total rejection region in half
- As such, non-directional hypotheses are more conservative compared to directional hypotheses when using the same alpha level
8
Q
Type I Error
A
- Alpha
- False positive
- “False alarm”
- Rejecting the null & concluding there is a relationship when there is in fact no true relationship
9
Q
Type II Error
A
- Beta
- False negative
- “Miss”
- Failing to reject the null & concluding there is no relationship when there is in fact a true relationship
(Finding no support for H_1 when H_1 is true)
10
Q
9 steps of hypothesis testing
A
- Specify your research hypothesis (and directionality)
- Formulate null and alternative hypotheses (including your research and/or statistical hypothesis for each)
- Specify your alpha (α); .05 is customary, but consider sample size as well [note that the timing of this step is non-negotiable].
- Select sample and implement design
A. Recruit participants, use existing admin data
B. Determine how constructs are measured
C. Collect data (if applicable)
D. Describe shape, distribution, patterns of missingness (such as MCAR or MAR) - Select appropriate parametric or non-parametric test
A. Decision based upon type of measurement and distribution of the data
B. Must ensure assumptions before running a test
C. Consider robustness - Compute test value (e.g., zobs, tobs)
- Find rejection region/critical value of statistic, based on sampling distribution of the statistic
- Check results against the critical value – if its absolute value is higher, your results are statistically significant.
- Reject or fail to reject null hypothesis
11
Q
Critical values
A
- Critical values are the specific numerical values that define the boundaries of the rejection region.
- This can be determined after choosing a significance level.
- On a two-tailed tests w a significance level of .025 on either end, the critical value is z = -1.96 and z = +1.96.
12
Q
Significance Level
A
- Significance levels determine how confidently we can test our null (and hence alternative) hypotheses.
- If the observed level of significance falls below our specified alpha, we can reject the null hypothesis because it is unlikely that the result is due to chance
13
Q
On a two-tailed tests w a significance level of .025 on either end, the critical value (z) is…?
A
z = -1.96 and z = +1.96.
14
Q
When do you use an Independent samples t-test?
A
- DV = continuous
- IV = binary (categorical)
- Additionally: independent observations w normal distribution and homogeneity of variance
15
Q
When do you use a dependent samples or paired t-test?
A
- DV = continuous
- IV = binary but paired category (e.g., time when scores on same measure at 2 time points)
- Additionally: diff b/w scores for each group or time point must be normally distributed (NOT independence of observations)
