Term
|
Definition
| Go beyond the sample to estimate, generalize to the population. |
|
|
Term
| Steps for Hypothesis Testing |
|
Definition
1. Identify the Research Hypothesis
2. Identify the Null Hypothesis
3. Assume the Null Hypothesis to derive a sampling distribution
4. Determine the Critical Region
5. Calculate the Sample Statistic for our sample
6. Make A Decision
7. Determine Chance of Error
|
|
|
Term
|
Definition
| Rejecting a true null hypothesis |
|
|
Term
|
Definition
| Failure to reject a false null hypothesis |
|
|
Term
|
Definition
| What we want to infer about the general population. It cannot be tested directly because we only have a sample. Might include direction eg. "positive correlation". |
|
|
Term
|
Definition
| Opposite of the research hypothesis. A hypothesis that assumes there is no correlation (or negative correlation in some cases) between the variables. It is used to derive a sampling distribution to compare with the research hypothesis. It may be univariate and gives a fixed value. |
|
|
Term
|
Definition
| A mathematical theoretical probability distribution based on what we expect to happen if the null hypothesis is true and we drew an infinite number of samples. |
|
|
Term
|
Definition
| Region of rejection. Extreme events that are so improbable under the null hypothesis that we reject it and infer the research hypothesis. |
|
|
Term
|
Definition
| The chance of making a type-1 error. Also called a probability value or p-value. It should be less than 5%. |
|
|
Term
|
Definition
| The critical region lies on only one tail of the standard normal curve and the hypothesis specifies a direction. |
|
|
Term
|
Definition
| The critical region lies on the left and right hand tails of a normal curve. |
|
|
Term
|
Definition
| A range identifying the probability that the sample statistic in the research hypothesis matches the population parameter. It is measured in units of standard error (z scores) or percentages and is usually set at 95%. This means that we are 95% confident we have captured the true population value in the interval. |
|
|
Term
| Explain what happens to a sampling distribution as sample size increases and why this occurs |
|
Definition
| As the sampling size increases, the sampling distribution becomes more accurate because we have a better estimate of the population parameter. We are less likely to get extreme outcomes if the null hypothesis is true |
|
|
Term
| What is meant by "proportional reduction in error?" |
|
Definition
| Proportional reduction in error _ how much better can we estimate the dependent variable (Y) using the independent variable (X) than estimating the dependent variable by itself? |
|
|
Term
| Measures of bivariate association |
|
Definition
| Lamda, Gamma, R and R2 (tau?) |
|
|
Term
| Lamda - Level of Measurement |
|
Definition
| 2 Nominal or 1 Nominal and 1 Ordinal; Lamda is a PRE |
|
|
Term
|
Definition
Variance = 0 (<) Lamda (<) 1.00
Lambda falls between 0 and 1 |
|
|
Term
|
Definition
| When mode lands in one row of dependent variable, Lamda will equal 0 |
|
|
Term
| Gamma - Level of Measurement |
|
Definition
2 Ordinal Variables; Measure of how much better we can estimate the order on the dependent variable knowing the order of the independent variable. Gamma is a PRE
|
|
|
Term
|
Definition
-1 < Gamma < 1
Gamma falls between -1 and 1
|
|
|
Term
|
Definition
| Assumes that high values in dependent variables will coincide with high values in independent variables and vice versa or lows with highs and vice versa |
|
|
Term
|
Definition
| Gamma fails to take ties into account, which inflates the correlation. Therefore Tau is used to take tied values into account (Tau is not a PRE). |
|
|
Term
|
Definition
| Tau corrects Gamma by including ties in the calculation. Tau is not a PRE and requires two ordinal variables. Absolute value of tau is always less than or equal to the absolute value of gamma. |
|
|
Term
| R and R 2 - Level of Measurement |
|
Definition
| Level of Measurement: Interval / Ratio; Use Ordinary Least Square Regression (OLS). R is not a PRE. R2 is a PRE |
|
|
Term
|
Definition
| R falls between -1 and +1; R2 is between 0 and 1; r gives the magnitude and the direction of the relationship. ASSUMES LINEAR RELATIONSHIP |
|
|
Term
|
Definition
| R and R2 assume a linear relationship between the variables |
|
|
Term
| What is meant by elaboration and control |
|
Definition
| Very few relationships are bivariate in nature so we use other variables (third, fourth, etc) to further explain (elaborate) and control the relationship between the dependent and independent variable. |
|
|
Term
| What 3 factors are appropriate to consider in control and elaboration? |
|
Definition
1-Demographics
2-Time and Place
3-Interest and Concern |
|
|
Term
| Why use elaboration and control (multivariate)? (2 reasons) |
|
Definition
a) to test for spuriousness
b) to elaborate on the relationship b/t x&y. |
|
|
Term
What can happen to the correlation when we add z? (3 things) |
|
Definition
a) nothing - z has no effect
b) correlation reduces drastically which means either the realtionship between x&y is spurious or z is an intervening variable
c) The correlation b/w x&y changes but does not disappear therefore z is a conditional variable |
|
|
Term
| Discuss the 4 major issues in questionnaire construction |
|
Definition
1-Content
2-Close vs Open Ended Questions
3-Response Format
4-Question Sequence |
|
|
Term
| What 8 things should you consider when you evaluate a poll |
|
Definition
1-Topic
2-Question Wording
3-Question Order
4-Sample Definition
5-Sub-group size
6-Error margins
7-Poll sponsor
8-What's the spin? |
|
|
Term
| Advantages to Interviewer Administered Survey Methodology (3) |
|
Definition
1-Higher response rate
2-Flexibility to probe/clarify
3-Quick results |
|
|
Term
| Disadvantages to Interviewer Administered Survey Methodology (3) |
|
Definition
1-High cost
2-Interviewer bias
3-Lack of anonymity |
|
|
Term
| Advantages to Respondent Administered Survey Methodology (4) |
|
Definition
1-Less expensive
2-No interviewer bias
3-Considered responses
4-Greater anonymity |
|
|
Term
| Disadvantages to Respondent Administered Survey Methodology (4) |
|
Definition
1-Low response rate
2-No one to probe/clarify
3-Takes time to get results
4-No control over who responds |
|
|
Term
| Advantages of Secondary analysis (3) |
|
Definition
1-Cost
2-May be the only source
3-Replicate results of primary data for accuracy |
|
|
Term
| Disdvantages of Secondary analysis (3) |
|
Definition
1-May only approximate data you want
2-May be biases you're unaware of
3-May be difficult to get access to data |
|
|
Term
|
Definition
| small groups (6 - 8 people) that evaluate something. It has advantages and disadvantages over public opinion surveys |
|
|
Term
| Focus Group methodology Advantages (5) |
|
Definition
1-Less expensive
2-Less prep time needed
3-No need to construct a survey
4-May not ask the appropriate questions on a survey (sort of fixed-in to set)
5-In-depth information |
|
|
Term
| Focus Group methodology Disadvantages (4) |
|
Definition
1-No external validity
2-Product is open to interpretation
3-May have moderator biases
4-Demand characteristics (do people behave differently) |
|
|
Term
| Two Major issues of ethics in conducting research |
|
Definition
| Informed Consent and Right to Privacy |
|
|
Term
| Ethics-Informed consent 3 issues |
|
Definition
1-They must be competent to make a decision (COMPETENT)
2-They must have full comprehension of the risks involved (COMPREHENSION)
3-The decision to participate must be voluntary (VOLUNTARY) |
|
|
Term
| Ethics-Right to privacy 5 issues |
|
Definition
1-Sensitivity of the information-how personal is the data? 2-Sensitivity of the collection location-some settings are inherently more private than others 3-Privacy of information-No identifiable information released 4-Anonymity-No one knows your name at any point during the test 5-Confidentiality-Individuals conducting the test know your name but do not release it to the public
|
|
|
