2 stats assignments - READ attachments FIRST
Instructions :
· Please answer each question below, placing the answer underneath each question (please do not do it in a research paper format, with a cover page, etc.).
· Answers should be at least two paragraphs each.
· Below is the link to the book (log-in info will be provided)
|
1. What is Significant? |
a. Define and apply the concept of “statistical significance” versus “practical significance” to something you are working on, or have read about.
b. Can you have one without the other? If so, name something that is statistically significant but not practically significant. How about something that is practically significant but not statistically significant?
c. Which is most important? Statistical or practical significance?
|
2. Contingency Tables (or Crosstabs) |
a. You will see contingency tables used often to compare two categorical variables (nominal and/or ordinal). To report results, you would simply compare the differences in percentages across categories. To establish if there is a statistical relationship (remember these are samples with margins of error), you could use the chi-square test. Do you have any contingency table examples to offer?
b. Why would you want to control for competing hypotheses (see the discussion in Chap. 16 of the textbook, Applied Statistics)? Give an example. Can you guess why multiple regression is the answer to this, in the long run?
[link to the textbook: https://www.vaultebooks.com/#/user/signin]
|
3. Comparing 2 Means and Chi-square Tests |
a. One of the most commonly used statistical techniques is the comparison of two means, also known as the t-test (for this you need quantitative variables, or interval level data to compare across two categories - find it on the RoadMap!).
b. The book spends a lot of time going over formulas for independent vs. paired samples. Don't focus on this. If in doubt, use the independent unequal variance formula. (It is the most conservative; that means if you can prove it with this formula, you can prove it with the others.)
c. T-tests (which is different from TDIST) are commonly used because they are the appropriate statistic for a comparison group or random assignment study (compare the mean from the treatment group to that of the control group).
d. If you are comparing two ordinal level variables, use a Chi-square Test. You might want to know if there is any relationship between being a Jets fan, categorized on a continuum from "not at all; green makes me sick", "am rooting for them now, I've seen the light (at the end of the tunnel?)", "rabid", "lifer" -- and whether or not you experience deep heartache, from "thankfully never -- I have no heart", "what is this new feeling and will it ever go away?", "every Sunday, unless they play another night", "every day of my life".
e. Can you cite a t-test or Chi-square result from a study, and describe its statistical significance (i.e., its p-value), and its practical significance?
f. Alternatively, take a difference of means or Chi-square problem from the book and show us how to do it, and how to interpret it. Ask questions about what is confusing.
Statistics Roadmap
Sheet1
| STATISTICS ROAD MAP | |||||||||||||||||||||
| Starting Point | |||||||||||||||||||||
| How many variables are involved? | |||||||||||||||||||||
| One Variable | Two Variables | Three or More Variables | |||||||||||||||||||
| What do I want to know? | |||||||||||||||||||||
| Central Tendency: | Dispersion: | Presence of Outliers: | |||||||||||||||||||
| Mean | Frequency Distribution | Boxplot | |||||||||||||||||||
| Median | Standard Deviation | ||||||||||||||||||||
| Mode | Range | ||||||||||||||||||||
| Charts and Graphs | |||||||||||||||||||||
| What is the measurement level of my 2 variables? | |||||||||||||||||||||
| Both categorical - nominal or ordinal | One categorical, the other interval | Both interval | |||||||||||||||||||
| What do I want to do? | |||||||||||||||||||||
| Hypothesis testing: | Strength & direction | ||||||||||||||||||||
| of relationship: | |||||||||||||||||||||
| Contingency Tables | Tau, Fisher's exact, | ||||||||||||||||||||
| Chi-square Test | sommer's d, gamma | Bivariate analysis, one categorical, the other continuous | |||||||||||||||||||
| Do I have 2 categories or more than 2? | |||||||||||||||||||||
| Categorical variable is dichotomous | 3 or more categories | ||||||||||||||||||||
| t-test of means | ANOVA | ||||||||||||||||||||
| Bivariate analysis, both are continuous | |||||||||||||||||||||
| Are the variables continuous or rankings? | |||||||||||||||||||||
| Both continuous: | Both rankings" | ||||||||||||||||||||
| Person's correlation | Spearman's rank correlation | ||||||||||||||||||||
| Simple regression | |||||||||||||||||||||
| Multivariate analysis | |||||||||||||||||||||
| What is my situation? | |||||||||||||||||||||
| One continuous | One dichotomous | Time Series: | More than 1 | ||||||||||||||||||
| dependent variable: | dependent variable: | Times series | dependent variable: | ||||||||||||||||||
| Multiple regression | Logistic regression | regression; IV | SEM | ||||||||||||||||||
| Source: Essential Statistics for Public Managers, Berman & Wang, 3rd Edition, 2012, pg. | MANOVA |