Research Methods Final Review PDF

# Research Methods Final Review ## Cumulative Concepts 1. **What is a dependent and independent variable?** (review from prior exams) - **Independent variable:** - Drives the study. - Variable of interest. - Reason for the study. - Variable that is manipulated. - **...

# Research Methods Final Review ## Cumulative Concepts 1. **What is a dependent and independent variable?** (review from prior exams) - **Independent variable:** - Drives the study. - Variable of interest. - Reason for the study. - Variable that is manipulated. - **Dependent Variable:** - Things being measured which are expected to change. - Change is due to the influence of the independent variable. 2. **Know the hierarchy of evidence pyramid** | Hierarchy of Research Designs & Levels of Scientific Evidence | | :------------------------------------------------------- | | Based on ability to control for bias and to demonstrate cause and effect in humans | | Clinical Practice Guidelines | | Systematic Reviews | | Meta-Analysis | | Randomized Controlled Trial | | Prospective, tests treatment | | Cohort Studies | | Prospective: cohort has been exposed to a risk. Observe for outcome of interest | | Case Control Studies | | Retrospective: subjects have the outcome of interest; looking for risk factor | | Case Report or Case Series | | Secondary, pre-appraised, or filtered Studies | | Primary Studies | | Observational Studies | | Narrative Reviews, Expert Opinions, Editorials | | Animal and Laboratory Studies | | No design | | Not involved w/ humans | 3. **Know the difference between a case report, case control, cohort, cross-sectional, descriptive, correlational, randomized controlled trial and meta-analysis.** - **Case report:** like a case study, but it may follow a few different cases, this is more than 1 person. - **Case control:** 2 groups, one group has the outcome variable (case) the other group does not (control). - **Cross sectional:** this is a snapshot in time, surveys are common, easy, quick, easy and cheap to collect. - **Descriptive:** goal is to describe features of a population; larger sample; surveys and questionnaires are common to gather data. - **Correlational:** can be used to make predictions; the stronger the correlation, the better the ability to predict (positive and negative correlation). - **Randomized control trial:** this is to help reduce bias when testing the effectiveness of new treatments or interventions. - **Meta-analysis:** A research process of examining a relatively large group of studies that have similar independent and dependent variables and determine a general strength of intervention for all of the studies. ## Developing a Research Question 1. **What is the difference between inductive and deductive reasoning?** - **Deductive Reasoning:** - Logical thinking that uses 1 or more general assumptions or theories to arrive at a specific answer or solution. - Starts with a general conclusion/ theories and then uses these to understand specific factors/ conclusions GENERAL → SPECIFIC. - Aim is to test a theory often with a hypothesis. - EX. (theory→ all knee injuries are painful. Observation→ trevor has injured his knee. Hypothesis→ trevors knee is painful) - **Inductive Reasoning:** - Logical argument in which one or more specific facts or observations lead to a general conclusion/ theory. - SPECIFIC → GENERAL. - Leads to conclusion/ theory; new theories are drawn from data 2. **What is the difference between primary and secondary sources?** - **Primary Sources:** Reports or documents provided directly by the person who authored it. - **Secondary Sources:** review of 1 or more studies presented by someone other than the original author. - 3. **What is the peer review process?** - Editors send out the manuscript to a handful of leading researchers in that topic area. - Articles are reviewed for suitability with a list of suggestions for revision (content changes; stats and methods issue; improvement in writing clarity). - Editor sends article back to authors with a letter telling them it will be accepted pending the changes. - Common to occur more than once. 4. **What is acceptance rate?** - The % of papers submitted that are eventually accepted for publication compared to rejected. - High quality journals may only accept 15-20% acceptance. ## Methodology 5. **What is the MAXICON principle?** - Maximize true variance - Minimize error variance 6. **What is the difference between a research population and research sample?** - **Research population:** entire group of individuals or elements that meet specific criteria and are of interest to the researcher. - **Research sample:** a subset of the research population 7. **What is the difference between inclusion and exclusion criteria?** - **Inclusion:** what subjects must have to be IN - **Exclusion:** what subjects must not have or they are out 8. **What is a power analysis used for?** - The greater the power of the study the greater the odds of arriving at statistically significant findings (the degree of confidence we have in the results of the study) 9. **What is the difference between random, convenience, stratified and systematic sampling?** - **Random sample:** selection procedure for participants in the study that provides an equal chance of selection. The best chance of representing the population if the sample size is large enough. - **Convenience sample:** a group of participants for studies selected because of convenient access. - **Stratified random sample:** appropriately represents subgroups do of the population. - **Systematic sampling:** every nth individual on a list is selected 10. **What is the difference between assumptions and delimitations?** - **Assumptions:** Things that must be trusted as true in the absence of actual verification, study cannot proceed without. - **Delimitations:** A limitation imposed by the researcher in the scope of the study 11. **What are the two general approaches to collecting data? Survey vs direct measure** - Survey vs direct measure 12. **What is the difference between open ended and closed ended questions?** - **Open ended**: - Do not have specific responses associated with them. - Respondents have an opportunity to form their own responses. - **Closed ended**: - More common. - Provide an embedded answer that the respondent can circle, check or list. - Answers are typically T/F, ABCD. 13. **What is a Likert scale?** - Researchers often use this as another question format. - Answers fall on a continuum which may have a scale of points that are equidistant apart. - (least stressed life 0-9 most stressed life) 14. **What is the difference between a questionnaire, interview, and focus group?** - **Questionnaire:** provides a good, quick methods of gathering data. - **Interview:** common in marketing and sales; keep it direct, short answers. - **Focus Group:** group interview (6-8 people is common); moderator guides discussion. 15. **What is the difference between a structured and semi-structured interview?** - **Structured:** All respondents are asked the same set of questions in the same order, regardless of response. - **Semi-structured:** the interviewer has a general framework of questions; however, the direction of the interview does sepend on the answers of the respondents 16. **What is measurement validity?** - Degree to which a particular measurement tool or method accurately captures the concept or variable it is intended to measure 17. **What are the differences between face, content, criterion and construct validity?** - **Face:** whether a test appears to measure what is claims. - **Content:** all relevant aspects of a concept are covered. - **Criterion:** compares test results to an external standard. - **Construct:** whether the test accurately captures the underlying theoretical concept being measure. 18. **What is measurement reliability?** - A measure of the repeatability of a test to get the same result when the same value is present 19. **What are three reasons which contribute to low stability reliability (poor reliability) of measurements?** - 1. Individuals perform differently from test to test - a. Sleep, injury anxiety. - b. Very hard to assess someone in pain. - 2. The measuring instrument may operate or be applied differently - a. Different clinicians. - b. Different education levels and background. - 3. The person administering the test may be different. - a. Physical abilities of a clinician. - b. Body type, strength, skill and practice. 20. **What is one reason for high stability reliability (high reliability) of measurements?** - Memory effect: Production of an artificially high score as subjects respond from their memory rather than the test itself. 21. **Why should we do pilot testing?** - Important to working out the kinks. - Preliminary experimentation done during the planning stages. - Before collecting “real data”. - Helps w determining DV & length of time. - See if measurements work. - Can help seeing variability of data. ## Data Analysis and Statistics 22. **What is the difference between nominal, ordinal, interval, and ratio data?** - **Nominal data:** - Does not have a numerical order. - Identified by name or label with no sense of order. - Frequency or percentages. - CANNOT calculate a mean or average. - EX→ food groups consumed in a meal; political party affiliation; occupation. - **Ordinal data:** - Order IS important. - Data that gives a ranking and meaningfulness of order - Scores have order but are NOT evenly spaced apart. - Can be reported as frequencies. - EX→ ranking by GPA; 1st or 2nd place; linkert scale. - **Interval data:** - True numeric data with adjacent numbers being equally spaced. - Distance between values is important. - 0 does NOT mean the quantity being measured is completely gone. - EX→ degrees of temp- 10 deg is 10 deg colder than 20 deg. - **Ratio data:** - The highest order, equally spaced numerical data - There is a true 0 where there is an absence of the quantity being measured. - EX→ age, weight, height, distance and time. - This is the most versatile type of data for statistical analysis. 23. **What is the difference between parametric and non-parametric data?** - **Parametric:** statistical tests used for interval and ratio data; must meet certain assumptions related to how data is distributed in the population it comes from. - **Nonparametric:** used for nominal and ordinal data; used when parametric assumptions are NOT met. 24. **What types of data collection techniques are used in descriptive statistical analysis?** 25. **What are frequency distributions?** - One of the simplest descriptive techniques. - Common in cross-sectional studies. - Often used with questionnaire formats for gathering data. - GOAL= simply determine the amount of times a variable occurs. 26. **What is mean, median, mode, range and standard deviation (do not worry about SEM)** - **Mean:** arithmetic average of a group of numbers (very popular) → sum of all #s, divided by the total number of values in the data set. - **Median:** always a single data value that resides in the middle of the data distribution; if there us an odd # of values the median is the center, if its even, the 2 central number are averaged to yield a single value. - **Mode:** least precise indicator of central tendency; the most frequent score in the distribution (MOST). - **Range:** difference between the high and the low scores: gives no info about the distribution of the scores. - **Standard deviation:** (most commonly used) serves as an estimate of the spread of scores away from the mean; the larger the SD, the greater the distribution of scores away from the mean. 27. **What is percentile rank?** - Goal is to indicate position of participants in a group based on scores. - Indicates one score position (ordinal data). - Commonly used when establishing or referencing norms. 28. **What is Pearson's r?** - Pearson product correlation coefficient. - The parametric statistical test for strength of relationship between 2 variables. 29. **What is the difference between a positive and negative correlation?** - **Positive:** one variable increases, the other increases (vise versa). - **Negative:** one variable increases the other descreases. 30. **What is considered a strong, moderate and weak correlation (know the r values).** - r= 0.8 and above= very strong. - r= 0.6 - 0.8= strong. - r= 0.4 - 0.6= moderate. - r= 0.2 - 0.4= weak. - r= 0.0 - 0.2= very weak/ no correlation. 31. **What is a null hypothesis?** - Statistical hypothesis states that no difference exists between groups tested in an experiment. - This is what we would expect of the IV had no effect on the DV. 32. **What is an alpha level?** - It is a “pre-determined p value” that you determine at the start of a study which is the “acceptable risk of making a type I error” you are willing to make. Usually set at P <.05 which means you will accept at most a 5% risk of making a type 1 error. 33. **What is a p value?** - The "actual P value” you calculate AFTER you collect data. This is compared to the alpha level you set at the beginning. If it is smaller, than your alpha level than you can reject your null hypothesis. If it is larger than you accept your null hypothesis (there is no difference between your study groups). 34. **What is a type 1 and type 2 error?** - **Type I (alpha- false positive):** rejecting a null hypothesis; stating there is a difference when there is actually no difference. - **Type II (beta- false negative):** accepting a null hypothesis; stating there is no difference, then there actually is a difference. 35. **What are the general difference between T-tests, ANOVA, Post hoc, ANCOVA, Chi square, Pearons r, Spearmans rho, Mann Whitney U and Kruskal Wallis ANOVA?** - **t-Test:** a statistical test used when comparing 2 groups on 1 DV - Type of data→ means from intervalv. - Groups→ 2; one group t-Test (population mean, sample mean, standard deviation and sample size), Independent t test for 2 groups, Dependent t test for 2 groups. - Comparison→ primarily used to determine whether there is a statistical difference between the means of groups, not to assess relationships. - Sample size→ small to moderate. - **Analysis of Variance (ANOVA):** - Data→ continuous (interval or ratio) as the DV; IV must be categorical. - Groups→ 3 or more groups; one-way ANOVA- 1 IV w multiple levels, two way ANOVA- 2 IVs examining main effects and interactions. - Comparison→ comparison of group means to determine a difference. - Sample size→ large. - **Post Hoc:** limitations of ANOVA tests; when 3 or more groups are compared using ANOVA, researchers,ust use this to determine which groups differe from each other. - Groups 3 or more. - Comparison→ comparing groups; determine which specific group means differ from each other agter ANOVA indicates a significant overall difference. - Sample size→ large or small. - **Analysis of Covariance (ANCOVA):** type of ANOVA that is used in cases where group means are known to be different→ 8 wk diff. - ^^^^^^^^^^^^^^^^^^^^^^ PARAMETRIC^^^^^^^^^^^^^^^^^^^^^^^^^^^^ - **Chi-Square:** used to compare frequencies or %s. - **One-way; two-way** - **Spearman's Rho:** non- parametric equivalent of the person product movement correlation r; used when variables are ordinal & sample size is too small. - **Mann-Whitney U:** non- parametric equivalent t test; compare 2 groups; uses with ranked data similar to Spearman's Rho and determines if the rank of one group is different than the rank of another group. - **Kruskal-Wallis ANOVA:** equivalent tp the ANOVA; used when 2 or more groups are compared; similar to MWU, data is ranked; interpretation is similar to ANOVA, payong heed to the p value for info about significance and strength off the group differences. ## Short Answer (you will be given a few of these exact questions below, I have included the correct responses for you to prepare) 1. **List the 9 steps important steps of a good methodology (do not worry about #10)** - 1. Pick a study design. - 2. Describe the population to be studied. - 3. Select the correct sample size. - 4. Determine your sampling process. - 5. Develop strategies to reduce bias. - 6. Create a consent form which participants must sign prior to data collection. - 7. Develop a protocol. - 8. Describe assumptions and delimitations. - 9. Select P values and statistical tests. 2. **What are the two types of experimental bias and three sources of error (effects) that occur in experimental research? What are 3 ways to reduces bias and error?** - 1. Sources of Bias and errors - 1. Sampling bias: - 1. bias that occurs when a sample is not representative of the population under study. - 2. Experimenter Bias: - 1. bias that occurs when the researcher unknowingly distorts the results of the study. - 3. Hawthorne Effect - 1. Subjects act differently because they know they are being watched. - 4. Placebo Effect - 1. Expectation of benefit from the independent variable, even if not benefit occurs. - 5. John Henry Effect - 1. Subjects intentionally act differently because they know they are not in the experimental group. - 2. First, we can reduce bias by randomizing sample selection and assignment to study groups. Next, we can have control groups with a placebo. Finally, we can blind participants, researchers, or both (double blind) 3. **When designing your “protocol” what are 8 key elements you want to take into consideration.** - 1. Collection of general demographic information (descriptive info). - 2. How many groups (experimental and controls). - 3. Length of your study. - 4. When will the dependent variable be measured? - 5. Important “operational definitions". - 6. Description of the apparatus or instruments used. - 7. Detailed explanation of what participants or investigators will be doing. - 8. Procedure should be written so it could be replicated. 4. **Why do we need statistics?** - 1. Statistics is a subfield of mathematics which helps us deal with the data we collect in a study. It includes aspects of collecting, analyzing, interpreting, and presenting data. - 2. It is the way researchers bring order to otherwise disordered data. - 3. It allows for comparisons to be made. Without statistics, we would not be able to judge the effectiveness of our independent variable. 5. **What are the two general purposes of statistics?** - 1. Description - 1. Determining how data may cluster together, be spread apart, or be related to each other in a data set. - 2. Inference - 1. Gives insight on how data is representative of a larger population, based on a smaller sample size. 6. **Before choosing a statistical test, you need to ask yourself what three questions?** - 1. What type of data do you have? - 1. Nominal, ordinal, interval, ratio. - 2. How may samples (groups) do you have? - 3. What is the purpose of the study (inferential, descriptive, comparison, relationships). 7. **What are 4 common types of scientific misconduct that we see in research?** - 1. Plagiarism. - 2. Poor data management - 1. Poor handling of outlier data. - 2. Falsification of data. - 3. Improper assignment of authorship. - 4. Multiple submission of work which does not allow for the peer review process. 8. **What are 4 common causes of scientific misconduct?** - 1. Pressure to publish. - 2. Need to complete graduate work. - 3. Need to get funding. - 4. Academic/professional promotion. 9. **What is the IRB?** - 1. In the United States, every college, university, hospital, and research Institute that engages in research on human participants must have an institutional Review Board. - 2. The IRB evaluates and reviews all research proposals to help insure the protection of human rights during research. 10. **What are the 6 key roles of an IRB?** - 1. The risk to participants is minimized with no unnecessary exposure and with use of sound research design. - 2. The risks to participants are reasonable relative to anticipated benefits. - 3. The IRB will be cognizant of vulnerable populations or educationally disadvantaged persons. - 4. Informed consent will be sought for each participant and a properly documented. - 5. Data collection processes are in place to ensure safety for participants. - 6. Participant privacy and confidentiality is protected. 11. **What is evidence based practice?** - 1. An integration of current research, clinical experience, patient values and the clinical setting.