PUBH 475 Week 13 – Relational (inferential) statistics Lectures

Vocabulary flashcards related to inferential statistics.

Inferential Statistics

Explore relationships between at least two variables to test your hypothesis, work with a subset (sample) to generalize about the population; examples include t-tests, ANOVA, Chi-square, and correlation.

Null Hypothesis (H0)

There is no difference between variables or groups; mathematically, the correlation and mean equals zero. No difference or relationship H0: M1=M2 or H0: M1-M2=0 No correlation: H0: rxy=0

Alternative Hypothesis (H1)

There is a relationship or difference between variables or groups; can be nondirectional, directional, or correlative. difference or relationships; H1: M1=/M2, M1-M2=0, Correlation: H1: rxy=/0

Scoring

The sum or mean score of all variables in a scale; may require reverse coding to ensure all variables are in the same direction; sums are used for inferential statistics instead of the individual questions. Turns to continuous variable

Revised 3 Levels of Measurement for Inferential Statistics

Nominal (categorical), ordinal (if <5, then categorical if >6, then continuous, and interval/ratio (continuous)

Inferential Statistic to run for a Categorical IV and a Categorical DV

Pearson’s Chi-square

Inferential Statistic to run for a Continuous IV and a Continuous DV

Pearson’s Correlation

Inferential Statistic to run for a 2-Level Categorical IV and a Continuous DV

Independent Samples T-test

Inferential Statistic for a 3 or More Categorical Level IV and a Continuous DV

ANOVA

Pearson’s Correlation Procedure

Examines the relationship between two continuous variables; test statistic is Pearson’s correlation coefficient (r); focuses on statistical significance, direction (positive or negative), and magnitude (0 to +- 1).

Independent Samples T-Test Procedure

Compares two categorical groups (IV) on a continuous dependent variable; test statistic is t; other t-tests include one sample and paired/dependent.

Analysis of Variance (ANOVA) or F-Test Procedure

Compares three or more categorical groups (IV) on a continuous dependent variable; test statistic is F; other F-tests include one sample and repeated measures.

Pearson’s Chi-Square Procedure

Examines the relationship between two nominal (categorical) variables; test statistic is x^2.

P-value

Probability value or level obtained in a test of significance; ranges from 0 (0%) to 1 (100%); used to determine whether to reject the null hypothesis; commonly used significance level is .05.

Procedure for Determining Significance

Run the test, examine the associated p-value, and make a decision to reject or fail to reject the null hypothesis; if p < .05, reject H0 and accept H1; if p >= .05, fail to reject H0; p between .05 and .10 indicates a trend towards significance.

Nominal Measurement

The weakest level of measurement, limiting the types of statistics that can be run.

Interval and Ratio Measurement

The strongest level of measurement, ranging from negative infinity to positive infinity, allowing several types of statistics to be run.

To determine which test to use we need to know the level of measurement for ?

Independent, Dependent, and each demographic variable