Sttatisics

*Class, z scores are not really very practical for dissertation research/quantitative methods. The MOST important value to identify and understand are P values (probability for Type I error). The gold standard for P values is .05 (5% chance for Type I error) or less. If an inferential (Pearson correlation, chi square, t test, ANOVA, etc.) or predictive (regression) statistical test yields a P value of .05 or less, the Ha (alternate hypothesis is supported and null hypothesis rejected). However, if a P value is >.05, the Ho (null hypothesis is supported and alternate hypothesis rejected). *See Hypothesis Writing and Testing examples!

*M3A1 modification (ignore the z statistic)

• Pair of hypotheses for each data point
• A p value for each data point
• Decision about the null hypothesis for each data point
• APA-formatted statement of results for each data point

Hypothessis Writing and Testing Examples

*M3A2 modification (substitute “P value” for “z score”) and answer the 3 related discussion questions.

RQ1. What is the relationship between length of time married and Marital Satisfaction as measured by the Marital Satisfaction Scale?

Ho1. There is no relationship between length of time married and Marital Satisfaction as measured by the Marital Satisfaction Scale.

Ha1. There is a statistically significant relationship between length of time married and Marital Satisfaction as measured by the Marital Satisfaction Scale.

*Key word “relationship” implies correlation!

RQ2. What is the difference between men and women’s Marital Satisfaction as measured by the Marital Satisfaction Scale?

Ho2. There is no difference between men and women’s Marital Satisfaction as measured by the Marital Satisfaction Scale.

Ha2. There is a statistically significant difference between men and women’s Marital Satisfaction as measured by the Marital Satisfaction Scale.

*Key word “difference” implies t test (independent t test if the groups of men and women are different in number or paired comparison t test if the groups of men and women are the same in number) for 2 variables (men/women). An ANOVA or MANOVA would be applicable if there were 3 or more groups (i.e., perhaps looking at marital satisfaction by ethnicity for example – African-American, Asian, Caucasian, etc.).

RQ3. What is the ability to predict Marital Satisfaction as measured by the Marital Satisfaction Scale?

Ho3. There is no ability to predict Marital Satisfaction as measured by the Marital Satisfaction Scale.

Ha3. There is a statistically significant ability to predict Marital Satisfaction as measured by the Marital Satisfaction Scale.

*Key word “predict” implies some type of regression (usually multi-nomial). For example, you would include all demographic variables in the predictive formula (e.g., age, time of marriage, gender, ethnicity, etc.). The Beta weights associated with each demographic variable would identify the strongest/weakest predictors of marital satisfaction in this case.