Research Method Quiz 9

3. What is the difference between a meta-analysis and a systematic review? What is the name for the number that you calculate for each study? What does it represent?

A difference between a meta analysis and a systematic review- a meta analysis is better. A systematic review is that you go and you have a topic like therapies for aphasia such as attention deficits and then you objectively do a search, and you list your search engines and search and search and you have a specific method for going through those articles that are applicable. When you find those that are applicable. Use a specific formula and find out the amount of gains in therapy- the effect size. You come up with a list of studies with an effect size and specify what the therapy method is. Both do that. A meta analysis goes a step further. If you have enough studies you can group them together. For example with agrammatism. We can have an average of all of those effect sizes. For a super effect size. As you do that you complicate things. To really public it you need a systematic way of evaluating the quality of the study.

4. What is the difference between a case study and a single subject design? Which is stronger? Why?

Case studies tend to have pre and post therapy measures. Single subject designs take a bunch of baseline measures and a bunch during therapy and a bunch after therapy. Single subject design is stronger because the subject serves as their own control group. They're used more often to see how effective interventions are. And much more measurements are taken.

**** More coming !!!!5. What does "degrees of freedom" mean? What affects the degrees of freedom? Do you want a high or a low number for degrees of freedom? Why?

Degrees of freedom for a statistic relate to the number of participants but reflect the number of data points that are independent or free to vary, or the number of observations that have no restrictions on their values. Degrees of freedom for a statistic are always the number of participants minus some number and for the pearson r correlation statistic, the degrees of freedom are the number of pairs minus 2. you want the number of participants to be high. So that the degrees of freedom will he higher so that you will be more likely to see a difference if there is one and not make a type 2 error. It will help you to see significance if there is significance."

Lecture quiz questions 1. What is the EBP triangle that I discussed? Describe the 3 difference points on the triangle. Explain what they mean clinically.

Evidence based practice triangle: best evidence at the top, clinical expertise and client/patient values at the bottom Best evidence is about using the best available information Clinical expertise : counts for something and should be considered. client/patient values: talk to the client and make sure that they are all right with your therapy methods and if they don't like something that you are doing, they have the right to ask for it to be changed and you have to respect their wishes. Talk to people about their goals - especially at the start of therapy.

5. What are the two lowest levels of evidence? Why are they considered low?

Evidence from a single descriptive or qualitative study and At the bottom is expert opinion. Expert opinion is considered low because there is no research to support it. Single descriptive or qualitative study is considered low because they do not take a pre and post measurement.

6. What is a Spearman rank order correlation? Is it parametric or nonparametric? Why would you use it instead of a Pearson correlation?

If the level of measurement is ordinal, researchers often use the spearman rank-order correlation coefficient (rho or rs) rather than the pearson r. It is nonparametric and it might be useful when your data does not meet assumptions associated with the use of the pearson r.

7. What is a chi square? What types of variables is it for? Give an example.

It is a procedure for investigating the degree of association between two categorical variables. a scientist wants to know if education level and marital status are related for all people in some country. He collects data on a simple random sample of n = 300 people, part of which are shown below. see phone for photo of table.

2. What type of research is considered the highest level of evidence? Why is it high? What type of research is the second-highest level of evidence?

Meta analysis is the highest level of evidence. It is high because it uses a statistical approach to combine the results from multiple studies in an effort to increase power (over individual studies), improve estimates of the size of the effect and/or to resolve uncertainty when reports disagree.The second highest level of evidence is Evidence obtained from at least one well-designed randomized controlled trial.

8. What is regression for? How is it different from correlation? What is the symbol for the regression statistic? Give an example of when you would perform a regression analysis.

Regression allows you to predict or estimate the value of the independent variable. A simple regression with two variables is similar to a correlation, except that one variable is the independent or predictor variable and the other is the dependent variable. See page 201 for the regression statistic (r2) = r squared One common use of regression analysis is to determine the value of the independent x variable for predicting future performance on the dependent y variable. An example of this would be research in the area of reading where researchers have investigated the predictive value of various measures of phonemic awareness to determine if they would provide early identification of children who would be at risk for reading problems in the future. In this example, measure of phonemic awareness would be the independent variables and measure of reading skills would be the dependent variables.

1. What does it mean to "reject the null hypothesis"? Is this something that researchers typically want to do? Why? Explain, using the terms "inferential statistics" and "Type...error" (fill in a number) why you can't prove anything with statistics.

To reject the null means that the probability of the hypothesis being true is low. You seldom will read a formal statement about rejecting or failing to reject the null hypothesis. Researchers are more likely to conclude that they found a statistically significant result or did not find a statistically significant result. Inferential statistics are a tool that helps researchers test their findings to establish how representative these finding are. In generalizing from observations of on a sample to the entire population the sample represents, researchers do not draw absolute conclusions. They do not make statements such as "out findings are or are not representative of the population in general." rather they make statements about how much confidence they have in their findings or about the probability of error associated with the results they reported.

2. Define Type I and Type II errors. Define them in relation to the null hypothesis, and then explain.

Type 1 occurs when a researcher rejects the null hypothesis when it is, in fact, a correct hypothesis. In other words, the researchers concluded that two groups were significantly different when in fact the two group were not different, or the researchers concluded that the correlation between two measures were significant, when the correlation was not actually significant. Inferential statistical tests are conservative and set up in such a way that the probability of making a type 1 error is kept low. Type 2 - the error occurs when a researcher fails to reject the null hypothesis when it is in fact an incorrect hypothesis. In other words, the researcher concluded that two groups were not significantly different when in fact the groups were different or the researchers concluded that the correlation between two measures were not significant, when the correlation was not actually significant. One reason that researchers might make a type 2 error is that the study lacked sufficient power.

3. What is the Pearson Product-Moment Correlation Coefficient? What does it tell you? (Talk about variables in your answer!) What is the typical letter that represents this measure?

Used for researchers who are investigating the relationship between two measures and the level of measurement is interval or ratio. The recommended symbol for the pearson correlation is r and this is called the pearson r. The potential values of the pearson r range from 0 to plus or minus 1. A measure of 0 would indicate the two sets of variables have no relationship, whereas a value of 1 would indicate the two sets of variables have a perfect relationship. The plus or minus indicates the negative or positive nature of the correlation.

4. What number corresponds to a weak, moderate, and strong correlation? What does a strong correlation say? What is a positive and what is a negative correlation? How are correlation results typically graphically represented?

Weak: 0.25 Moderate: 0.50 Strong: 0.75 and above A strong correlation says Correlation is a term that refers to the strength of a relationship between two variables where a strong, or high, correlation means that two or more variables have a strong relationship with each other while a weak or low correlation means that the variables are hardly related. Positive and negative: indicates the direction of the relationship They are typically represented with a scatterplot on a graph see page 195 for examples

6. List and explain 4 signs that a presentation of a "scientific" therapy procedure is baloney.

•Don't be fooled by testimonials - can get people to say anything. Not objective evidence. •Don't be fooled by their own research- like schwartz (stuttering solved) doing his not peer reviewed research so others don't take a look at it. •Does their device/product relate to established clinical practice? New things should relate to things that have already been done. Have studies been done in peer-reviewed journals? Martin F. Schwartz never published his stutter no more book in a journal, because then it would be peer reviewed. Don't trust things that just come up in books. Good thing about peer reviewed is people can respond to it and research it.