Description

Several statistical doctoral students decided to create a special 6 week stats online tutorial, STATWORLD, for high school students. They convinced Wake and Durham Counties to allow them to determine if STATWORLD would increase students’ scores on the AP Statistics Exam. The creators randomly assigned 100 students to take STATWORLD, before they took their stats course and another 100 students would serve as a control group.

Variables

STATWLD take the online tutorial or not; where 0 = no and 1 = yes

PreSTAT pre-test b/f taking the high school statistics course ( 0 -100 pts.)

PostSTAT post-test a/f taking the high school statistics course (0 -100 pts.)

Sgender male or female (1 = male & 2 = female)

SSES students socioeconomic status group (1 = low; 2=medium; 3 = high)

SchSES school’s overall socioeconomic rating (0 = less than 40% receive free/reduced lunch; 1 = otherwise)

Tcertif teacher is licensed in the state to teach mathematics (0= no license; 1 = license)

Tprofdev teacher has received professional development coursework for STATWLD (1 = none; 2 = basic; 3 = advanced)

Research Design (2 points total)

1. What type of design is this research project? (1)
2. Are there any limitations to this design? If so, what are they? (1)
   Descriptive Statistics (24 points total)
3. What is the scale of measurement for each variable? (4)
4. What is the most appropriate measure of centrality for each variable? Compute it. (4)
5. What is the most appropriate measure of variability/dispersion for each variable? Compute it. (4)
6. Graph each variable and describe the results in 1 – 2 sentences. (4)
7. Describe student #27 for the 8 variables. (4)
8. Describe the data set in a paragraph. (4)
   Inferential Statistics (bivariate relationships - 56 points total )
9. Examine the bivariate relationships among the variables (i.e., 28 relationships). For

each bivariate relationship, state:

1. The type of test used (independent t-test, Pearson’s r; Spearman’s rho; ANOVA with
   or without post hoc, and chi-square) (14 points)
2. The null hypothesis in symbols and a sentence. (14 points)
3. State the results in an APA sentence , which generally includes the (1) test statistic (e.g., F, r, t, etc.), (2) sample size, (3) degrees of freedom, and (4) p-value. (14 points)
4. Give an English translation of the results of each test. (14 points)
5. Briefly, from the bivariate results, which variables are significantly correlated with the dependent variable (i.e., post stats score).

Multiple Regression -( 43 points total )

10. Theory has indicated that student characteristics, teacher characteristics, and school characteristics account for statistic achievement scores. As such, the researchers are going to run 4 regression models to determine to what extent these factors predict achievement in statistics. (Special note; a mixed-model and HLM designs are more appropriate, but they are beyond the scope of this task. As such, we will use OLS regression)

Remember that you will need to create dummy codes where appropriate. In the case of selecting a reference group, use: (a) theory, (b) convention, and (c) the middle value to guide your decision.

The dependent variable for all models is post-test achievement on statistics exam (POSTSTAT) .

**See APA table at the end of assignment to help with reporting data in APA format.** Use Green and Salkind otherwise.

1. Regression Model 1-Student level

Predictors: Pre-test, gender, STATWORLD, and student’s SES

• Create an APA formatted table and insert results. Explain the results in APA style. (5 points)
  B. Regression Model 2-Teacher level
  Predictors: STAT WORLD, certification, and professional development
• Create an APA formatted table (include b and β) and insert results. Explain the results in APA style. (5 points)
  C. Regression Model 3-School level
  Predictor: STATWORLD and School SES
• Create an APA formatted table and insert results. Explain the results in APA style. (5 points)
  D. Which model was a better predictor of statistics achievement? How do you know? (2 points)
  E. Now, put all of the variables (student, teacher, school-level, and STATWORLD) into one model.
• Create an APA formatted table and insert results. Explain the results in APA style. (5 points)
• Did any of the variables become significant or nonsignficant? (compared to earlier models) Explain. (2 points)
• Write the equation (Greek letters and b coefficients) for this model. (2 points)
• What does b (i.e., unstandardized coefficient) for the pre-test mean for this data set? (1 point)
• Describe what the constant represents for this data set. (2 points)

11. Let’s test our assumptions (6 points)

1. Test for linearity between pre and post statistics achievement. (use a scatterplot). Is the relationship linear? Explain.
2. Test for homoskedasticity (residuals plot—pre and post statistics achievement). What do the results indicate (homoskedasticity or heteroskedasticity)? Explain.
3. Is POSTSTAT normally distributed? How do you know?

12. Ignoring any violations of the assumptions, from this data set, state: (8 points)

1. Brief summary of the results (few sentences)
2. Limitations of the study (at least 2)
3. Practical implications (at least 2)—what would you tell school or district-level officials considering purchasing the STATWORLD online tutorial?
4. Future research directions for the STATWORLD research team (at least 2)—what should they do next; any unanswered questions; any problematic aspects of the data; can we redesign to overcome the study’s limitations?

Price: $49.99

Solution: The downloadable solution consists of 34 pages, 2689 words and 59 charts.
Deliverable: Word Document