prefinal reviewer

• Hypothesis testing


  The process of testing our hypothesis to ensure that our explanation or hypothesis about a behavior, event, or phenomenon is plausible - reasonable, and supported by evidence
• Steps of hypothesis testing

  1. State the research question

  2. Specify the null and alternative hypothesis

  3. Determine the level of significance

  4. Compute the test statistic

  5. Make a decision and state conclusion
• Research question


  The foundation of a systematic investigation, representing an inquiry the research aims to answer. It is written in question form and has no stated expectation about the outcome.
• Types of research questions

  • Testing of Difference: Do men and women differ in terms of conversational memory?
  • Testing of Association: Is there a significant relationship between social media use and level of self-esteem among adolescents?
• Hypothesis


  A guess, possible state, or tentative answer to the research question, based on research and grounded in theory
• Rationale of hypotheses


  The logical reasoning behind the direction of the hypotheses, which may be based on past research, existing theory, or logical reasoning
• Null hypothesis


  A prediction of no effect on treatment, no relation between variables, and no difference between groups. The main idea is that nothing is going on.
• Alternative hypothesis


  A prediction of real effect on treatment, the relationship between variables, and the difference between groups. Can be directional (one-tailed) or non-directional (two-tailed).
• Level of significance


  The level of probability at which the null hypothesis can be rejected with confidence and the research hypothesis can be accepted with confidence. Conventional level is α = .05, conservative level is α = .01.
• Parametric tests


  • Assume normality (or at least large sample sizes so that sampling distribution is normal), require interval or ratio level of data, and are the most powerful
• Non-parametric tests


  • Do not require a normal distribution, can be used in both nominal or ordinal levels of measurement, and are utilized in conditions that are not stringent or controlled, or when the sample is small (n<30)
• Parametric tests

  • Independent Samples t-test
  • Paired Samples t-test
  • Analysis of Variance (ANOVA) - One Way or Two Way
  • Pearson Correlation
  • Linear Regression
• Non-parametric tests

  • Chi-Square
  • Kruskal-Wallis H
  • Wilcoxon Test
  • Friedman Test
• Decision-making in hypothesis testing

  1. Compare the p-value with the significance level

  2. If p-value < significance level, reject the null hypothesis (statistically significant)

  3. If p-value ≥ significance level, fail to reject the null hypothesis (not statistically significant)
• Conclusion in hypothesis testing

  1. Describe the results of the null hypothesis (whether it was accepted or rejected, and for what value of alpha or p-value)

  2. Describe the results of the alternative hypothesis (answering the research question and stating the sample statistic collected)
• Type I error (α error)


  Rejecting the null hypothesis when in reality the null hypothesis is true - finding a significant effect, difference, or relationship, when in reality, there is none (false positive)
• Type II error (β error)


  Failing to reject the null hypothesis when in reality null hypothesis is false - finding no significant effect, difference, or relationship when in reality, there is a difference (false negative)
• Significance level


  Regulates Type I error - conservative standards reduce Type I error, but increase the probability of Type II error
• Sample size


  Regulates Type II error - the larger the sample, the lower the probability of Type II error occurring in conservative testing
• Inferential statistics


  A set of tools and procedures wherein the data are used to estimate population parameters based on sample values or to test hypotheses and draw conclusions
• Non-parametric alternatives to parametric tests

  • Mann Whitney U or Wilcoxon Rank Sums Test (alternative to Independent Samples T-test)
  • Wilcoxon Signed Rank Test (alternative to paired samples t-test)
  • Kruskal Wallis H (alternative to One-way and Two-way/Factorial Anova)
  • Friedman Test (alternative to Repeated Measures ANOVA)
  • Spearman's Rank Correlation (alternative to Person's Correlation)
• Types of ANOVA designs

  • One-Way ANOVA - compare more than two groups/levels
  • Factorial ANOVA - use more than two independent variables
  • Repeated Measures ANOVA - measure each respondent three or more times/conditions to look for change
• One-Way ANOVA


  A statistical test and non-directional procedure that compares more than two means to determine if they have statistically significant differences
• Logic of ANOVA


  • Variance - a measure of how spread out or dispersed the scores are from the mean
  • Two sources of variance: Variation Between Groups and Variation Within Groups
  • Comparing means using ANOVA entails an analysis of whether the distribution of groups are separate enough
• Assumptions of One-Way ANOVA

  • The dependent variable must be continuous (interval/ratio)
  • The independent variable should consist of two or more categorical independent groups
  • All observations are independent of one another
  • The dependent variable is approximately normally distributed for each group of independent variables
  • No significant outliers
  • There needs to be homogeneity of variances
• Concepts of ANOVA


  • Between Groups Sum of Squares: Deviation of each group mean from the total/combined mean
  • Within Groups Sum of Squares: Deviation of raw score from the group mean; Similar to variance
  • Total Sum of Squares: Deviation of raw score from the total/combined mean
  • df: degrees of freedom; k= number of groups; N - total number of sample size
  • MS: Mean Squares - a measure of variation that controls for the number of scores involved
• Significance and effect size

  A significant difference simply indicates that the observed statistic is unlikely to occur by chance or sampling error. To find out how large the difference/effect is, we compute for the EFFECT SIZE.
• Post-hoc test


  Used to determine exactly where the significant difference(s) lie after a significant ANOVA result
• Standard error


  The quantification of the sampling error, measuring the variability or dispersion of the sample mean from the population mean, used to estimate the population of the mean and estimate differences between population means
• Effect size


  The value that shows the practical significance of the statistical results, giving researchers an idea of how large, important, or meaningful a significant effect is (0.2 - small, 0.5 - moderate, 0.8 - large)
• 1. test table


  Used to find the t-critical value to decide whether or not the null hypothesis should be rejected
• Degrees of freedom


  The factor that made sample standard deviation differ with population standard deviation
• Confidence interval


  A range of values within which the true population parameter is estimated to lie with certain level of confidence. If the range brackets the null hypothesis value, we fail to reject the null hypothesis. If it does not, we reject the null hypothesis.