CDIS 3413 1st EditionFinal Study Guide- Null Hypothesis: How things are without making any changes; no differences, no effecto Ex: If you were testing whether or not a treatment caused any improvement in a client, the statement, “The treatment will have no effect on the client,” would be thenull hypothesis.- Alternative Hypothesis: What you are trying to prove; the expected resulto Ex: In the previous example, “The treatment will improve the client’s language,” could be an alternative hypothesis.- Type I error: Rejecting the null hypothesis (H0) when it is actually correct.o α represents the possibility of type I error. If α=.5, then there is a 5% chance that the researcher has made a type I error. Typically α=.5- Type II error: Failing to reject the null hypothesis when it is actually incorrect (rejecting the alternative hypothesis when it is correct)o β represents the possibility of type II error. If β=.2, then there is a 2% chance that the researcher made a type II error.- One-tailed hypothesis test: typically worded like the previous example; a treatment either helps or does not help, etc.o- Two-tailed hypothesis test: the research question will allow for a treatment to cause changein either direction (improve or get worse, mlu increase or decrease, etc.)o- Empirical Rule: o On a normal curve, the further away from the mean you get, the lower the probability of someone scoring that score. On the graph below, the line labeled 0 is the mean.o 68% of scores are within ± 1 SD of meano 95% of scores are within ± 2 SD of meano 99% of scores are within ± 3 SD of meanRejection regiono- Skew (kurtosis): the measurement of peakedness. o Skewed distributions are not symmetricalo Positive (right) skew: Tail is on the right side. Most of the data points are on the left.o Negative (left) skew:  Tail is on the left side. Most of the data points are on the left.- Parts of a scientific papero Introduction should: Introduce the problem Develop the background State the purpose and rationaleo Method describes: Participants Apparatus and materials Procedures - Should have much detail to that the research can be duplicatedo Results should: Summarize the data collected and the statistics or data analysis used Include tables and figures to display results Report statistical significance, confidence intervals, and effect sizeo Discussion should: Evaluate and interpret the implications of the results Include a statement of support or nonsupport for the research thesis Comment of the importance of findingso Title and Abstract should: Be written last Name the variables Be concise Indicate what was studied, but not results or conclusions Mention the population of interest Use subtitles to demonstrate the type of research- Properties of z and t distributionso z and t scores are test statistics.o The scores indicate differences between sample statistics in standard error units.o Which one you use depends on the sample size.o Comparing test statistic to the critical value: If the test statistic’s absolute value exceeds the critical value, then reject H0. If the test statistic’s absolute value does not exceed the critical value, then failto reject H0.oz=x−´xs- 5 phase model of clinical research (the goal is organization of research)o Phase I: explorative Detect positive or negative treatment effectso Phase II: explorative Further refine definitions, population, and methodso Phase III:  Test specific hypotheses regarding treatment efficacy in a larger sample and more rigorous study.o Phase IV: bridge between research and practice Establish effectiveness of treatment in real world settingso Phase V: other effectiveness issues Cost-benefit analysis, consumer satisfaction, quality of life issues- Efficacy: how well a treatment works during clinical trials in the laboratory- Effectiveness: how well a treatment works in real-life settings- Benefits of synthesized research reports:o Systematic review: Narrative Approach Reduces reviewer bias Thorough search of pertinent literature Qualitative analysis of previous findings Conclusion based on synthesis of findingso Meta-analysis: Quantitative Approach Combines effect size measures from individual studies Presents an overall measure of effect Tests significance of overall effect- Descriptive statistics: “description” of sample with data at hand- Inferential statistics: make “inferences” about population based on sample- P-value: the probability of getting a test statistic as extreme as the one you obtained if H0 were actually true.o Smaller p-values are better.o If p<.05 you typically reject H0.o Alpha values are set before the trial is conducted. P-values are obtained from the data after the trial.- Cohen’s do Difference between means in standard deviation unitso d = =varianceCalculating standard deviation:1. Find the mean2. Subtract the mean from the value of each data point3. Square the result4. Sum those results5. Divide the sum by (the number of data points, n, minus one)6. Now that you have the variance, take the square root of that number to find the standard deviation.M1 −