Download Data Analysis Exam Study Guide: T-tests and ANOVA and more Exams Data Mining in PDF only on Docsity!
DATA ANALYSIS EXAM 2 EDITED VERSION 2025
"The scale of measurement is met because the dependent variable (X) is on a continuous scale" normality - look at histogram to see if normally distributed and skewness/kurtosis should be between - 1 and + "The requirement for normality is met because both skewness and kurtosis range from - 1 to +1 (Skewness: x.xx, Kurtosis: x.xx). The graph also reveals a normal histogram with 1 peak." outliers - look at boxplot "There are no outliers which meets the requirement." equality of variance - look at Levene's test: want it to be NOT significant, so p greater than .05. If not, use Welsch correction table value. "The assumption for equality of variance is met because the Levene's test is not significant (p= .xx)" one sample t-test example research questions - Compares the mean of sample to a hypothesized value
- Does our sample have a developmental score that's greater than 15?
- IS GPA from this class different from average GPA of 3.11?
- Does sample have a longer time than average mile time of 8 minutes?
paired sample t-test example research questions - Natural pairs of scores (pre/post, husbands/wives)
- Do couples differ in cleaning score?
- Did your motivation significantly increase after receiving treatment?
- Do twins differ in motivation? independent sample t-test example research questions - Compares means of 2 different groups of scores to each other
- How do married and single women differ in life satisfaction?
- Do males and females differ in motivation?
- Do freshman and seniors differ in IQ scores? P-values significance - Divide by 2 for one-tailed test (directional hypothesis) p > .05 = not significant p < .05 = significant SPSS t-tests - 1. Check assumptions (Analyze- Descrip. stats- Frequencies)
- Run Analysis (Analyze-Compare means- type of test)
- For one-sample t-test: Change test value
- For paired sample: Move both variables (pre/post)
- For independent sample: Move independent variable to "Grouping variable box" and dependent to "test variable" and hit define groups
scale of measurement - dependent variable must be on a continuous scale (interval or ratio) ANOVA - Can handle more than 2 groups on one outcome variable one-way ANOVA example research questions - Multiple groups
- Do catholics, traditional, and charismatic christians differ in level of perceived spirituality?
- Do the two different experimental groups and control groups differ in anxiety levels? steps for completing a one-way ANOVA - 1. Make sure a one-way ANOVA is the correct analysis for your research question
- Check ALL assumptions- one-way ANOVA includes EOV (Levene's test)
- Look at the overall one-way ANOVA output (look at p-value)
- Run post hoc analysis if necessary (if p-value is significant) types of post hoc analyses - Scheffe (most power) Tukey Bonferroni (least power) Format for writing results for significance for t-tests - t(df)= t, p-value ex. t(27)= - 7.13, p<.
Format for writing results for significance for one-way ANOVA - F (Between groups df, within groups df)= F value, p-value ex. F (3, 11)= 9.00, p=. Format for writing results for significance for factorial ANOVA - F (interaction df, error df) = F, p-value If p is .000 - write as p <. 001 SPSS One-way ANOVA - 1. Run assumptions test (Analyze- Descrip. stats- frequencies)
- Run analysis (Analyze- compare means- one way ANOVA)
- Put dependent variable in "Dependent list" and independent in "factor"
- Under "options" click descriptives, homogeneity of variance test, welch if levene's test is significant
- Under post hoc- select one type Factorial ANOVA - how we get an interaction; shows how dependent variable is impacted while looking at the independent variable at the same time Factorial ANOVA example research questions - Multiple variables
- Does gender and political affiliation have an impact on political knowledge?
- Does gender and ethnicity have an impact on test scores?
and kurtosis values are in the range from - 1 to +1 (Skewness: - .902, Kurtosis: .402). Additionally, the histogram was normally distributed. The analysis showed that the result was significant; t (27) = - 5.148, p < .001. Specifically, our sample had a significantly lower aggression level compared to the national average of 54. paired sample t-test format long answer - A paired sample t-test was conducted to determine if there was an increase in aggression levels after given food coloring. (See format for assumptions). The analysis showed that there is a significant difference between pre and post levels of aggression; t (27)= - 7.343, p < .001 (one tailed). Specifically, aggression levels increased after food coloring was added compared to when there was no food coloring. independent sample t-test format long answer - An independent sample t- test was conducted to determine if males and females have different pre levels of aggression. The assumption for equality of variance is met because the Levene's test is not significant (p= .061). (See format for other assumptions). The analysis showed that there was not a significant difference in the participants levels of aggression; t (26) = .693, p = .501. one-way ANOVA format long answer - A One-way ANOVA was conducted to determine the differences between ethnic groups and income. Before the analysis was conducted, assumptions were addressed. (See format in previous examples). The assumption for equality of variance is met because the Levene's test is not significant (p = .301). The results from the one-way ANOVA were significant; F (3, 95.399) = 30.059, p < .001. Because the results were significant, a Tukey post hoc analysis was conducted. This was chosen to account for some power but not too much. The results show that Asians and Caucasians have a higher math total score compared to Hispanics and Blacks, however Caucasians do not have a higher score compared to Asians.
factorial ANOVA format long answer - A factorial ANOVA was conducted to determine if a salesperson's performance is influenced by their gender and type of training received. (See format for assumptions). Results from the factorial ANOVA were significant from the interaction between group and gender;; F (2, 54) = 6.490, p=.003. Because the interaction was significant, the interaction plot is looked at. The plot reveals that females have higher performance scores unless they are in the third sales training group, in which case males have a higher score.
