Research Methods Unit #2
Replication Designs: Conceptual
Conducted to establish validity and generality of particular concepts or theories Usually based on same literature review and test the same hypotheses, but use different methods Referred to as derived studies
Writing the Introduction
Should introduce the problem Develop the background State purpose and rationale Tell readers what article or report is about in an interesting manner Asking questions, present hypotheses
Two types:
Simple Case Studies Time Series Design
Overview Cont.
Some research results go unpublished or are not available for review, mostly due to: Editorial rejections: occur because of lack of theoretical significance or some limitation in study's methodology Editorial censorship: controversial issues Self-censorship: researchers choose to not report their publishable findings Writing is more than just dissemination of the research product; it helps generate new ideas and rethink old ideas Goals of writing: Inform: present relevant facts Explain relevant facts Interrelate relevant facts Persuade audience of authors explanation
Efficacy of treatment in Single Case Design
Statistical tests for single case designs: Data points are NOT independent - they are related to each other Thus, cannot use analysis of variance (ANOVA), t-tests that assume that data for analysis is independent Can use interrupted time series analysis (ITSA) which examine overall change in level of performance or change in slope of graph Also, use of effect sizes (ES): measure of magnitude of change in behavior ***ITSA-change of the level of performance. makes a decision whether the change in slope is significant. change in the slope of the graph. Using effect sizes-the size of change from baseline to criteria measure of the magnitude in the change of behavior. overall, any change that is 1.00 and over is considered a remarkable or significant change in behavior. .82? still okay. .11? that would not be remarkable or significant. approaches used to look at single subject design.
EBP: Identify Problem, gather information
Two sources of information: Raw evidence: information not subjected to expert review Pre-filtered evidence: information reviewed by experts (peer-reviewed journal articles) Research in scientific journals offers more confidence in scientific credibility; however, does not mean findings are truth Searching databases for pertinent information: need appropriate databases and key words. ****Articles in the ASHA leader is a periodical...in Scientific journal it is a data based study and it is peer reviewed. It will have references but most likely not a data based study.
Simultaneous Treatment Design
Design may include additional treatments but difficult to balance more than two treatments at a time Multiple treatment interference effects are possible as threat to internal validity Thus design requires careful implementation to get valid results
Grounded research
Develop theories about phenomena of interest that are grounded in observation Verification of information Observe, verify, and iterate until researcher is satisfied that data are sufficient to develop credible theory about phenomenon of interest ***Phenomenology type research or view point.
How does the study fit into what is already know?
Does study provide new knowledge? Does it test a new program? Does it contribute to what we know about the target problem? Does it contribute to what we do NOT know about the target problem
Qualitative research approach cont.
Dynamic systems: dynamic process in which answers, questions, theories may change throughout course of study Unique case orientation: each case is unique and deserves detailed, in-depth study Context sensitivity: appreciating the contextual variables that influence the phenomenon of interest Empathic neutrality: Qualitative researchers are neutral and non-judgmental observers Design flexibility: redirection of focus, generation of new questions, as they emerge during course of study
Evidence-based Practice, and Information Literacy
EBP involves systematic use of best evidence for solving clinical problems; includes 8 steps: Identify clinical problem Gather information from research studies about problem Ensure you have adequate knowledge to read and critically analyze research studies Decide if a research article/review is relevant to clinical problem Summarize information for use in clinical practice Define expected outcomes for clients/families Provide education/training to implement change in clinical practice Evaluate change and modify if needed ***Steps to take when you take on a new client.
Alternative Methods in Research Design (Overview)
Not all research endeavors aim to evaluate cause and effect Some seek to describe participant behaviors with no attempt to change or manipulate behaviors These designs include: case studies, ethnographic research, historical research, correlational research, survey research
Variations of ABA/ABAB Single Case Design
ABA and ABAB designs typically include one IV (treatment) to test effect of intervention Researchers may want to compare effects of two treatments: can use ABAC design First treatment is B, second treatment is C with a baseline phase between them Need to be concerned with multiple treatment interference effects Can control by using more participants and counterbalancing order of treatments across participants ***have to be concerned with multi-treatment effect. they do this because they want to compare two different treatments only way you can do if you take several subjects with 3 receiving C treatment first and three receiving the B treatment first. After second treatment, you are still dealing with the combined effects of the two treatements to see if we can make a diff in the individuals, not comparing treatment at this point, example: indiv with ASD try different treatments, just interested in the person getting better and not comparing 2 treatments.
ABA/ABAB Single Case Design
ABA is prototype for this type of research ABAB is an extension of ABA adding a B phase to the sequence ABA indicates a series of three phases implemented over time First A phase is initial baseline phase (DV is observed in its natural state) B represents intervention phase where treatment is introduced Second A is second baseline phase in which treatment is withdrawn
Developmental Research in CSD
Designs repeatedly measure behaviors over a period of time in order to record developmental trends The independent variable (IV) is typically maturation; the dependent variable (DV) is whatever behavior/trait researcher chooses to study Maturation can be operationally defined as chronological age; however, it is underlying attributes of physical and mental changes that take place over time
Qualitative Research Designs: Direct Observation Method
Detailed and systematic observations of person, location, event, topic without researcher's participation in scene Reduces researcher bias Utilizes note taking, video/audiotaping, photographs Usually data collection is triangulated from different sources to insure credibility of observations Use of one-way mirrors, video cameras ***you can not be observed. Your presence can affect behavior.
Description of participants?
Detailed description permits credible evaluation of results Allows practitioners to judge if the results can be transferred to their clinical settings Allows other researchers to replicate study for result confirmation Permits other researchers to perform secondary data analysis
Qualitative research approach
Alternative to quantitative methods when individual perspectives are more important than group generalizations When primary purpose is exploratory Generating theories as opposed to testing hypotheses All approaches to qualitative research identifies role of researcher, stages of research process, and methods of data analysis (The role of the researcher is very important in gather and analysis of the data. (you are more of an examiner in quantitative) stages important here. Its important to know as a researcher what stage you are at. )
What is a survey?
Any measurement procedure that asks questions of respondents taking form of questionnaires or interview instruments Collect data on sociological or demographic variables and/or psychological variables such as opinions, attitudes Questionnaires: usually mailed and economical and convenient Limitations: response rate is poor; not reliable instruments for gathering detailed responses from participants Validity: researchers must carefully choose content, wording, format, and placement of questions in surveys ***Even placement of the questions within the survey is important. And wording. Incentive. Balance-people could take offense to a non-offensive question.
Description of Procedure
Are participant instructions clearly described? Is it clear what the dependent variable was and how it was measured? Timeline for collecting data described? Are manipulations of the independent variable(s) clearly described? Are operational definitions for variables provided? Is study controlled to eliminate influence of extraneous variables?
Equipment/Apparatus Stimuli
Are stimuli used clearly described? Is each instrument, piece of equipment used clearly described? Can you determine if instruments are reliable and valid?
Multiple Baseline Design
At this time, treatment is discontinued for target variable 1 because criterion is reached. Maintenance data is continued to be collected on target variable 1 (examine degree of carry over) while treating target variable 2 Design demonstrates treatment effect by showing response changes whenever treatment is introduced at different points in time. ***only show improvement when treatment is introduced on the behavior you are working on.
Rewriting and Revision
Authors typically encounter some serious revisions following initial writing and first draft Hard to edit your own writing Requires a high degree of attention to detail Substantial restructuring of paper is sometimes necessary
Publication/presentation Stage
Before submitting for publication, possibly ask colleague to critically read paper, provide constructive criticism No research product passes review without some revisions Authors will have opportunity to revise paper if it is not rejected Need to carefully address reviewers' comments Should submit cover letter with revised paper explaining how revision addresses and includes reviewers' comments and suggestions. If authors do not incorporate reviewer suggestions, they need to explain their rationale This process continues until research paper is ready for publication
Baseline in Single Case Design
Begin with systematic observations of an individual's behavior for several sessions before treatment is introduced Known as baseline phase or pre-treatment phase Purpose: Identify/describe extent of person's problem (status of target behavior in its naturally occurring existing state) Predict future behavior if intervention is NOT provided ***what to do at least two or three baselines. the baseline identifies the persons problem. how it right this minute before I introduce the treatment. this is predicting future behavior if we didnt intervene. Example: mr. smith retrieving words in a sentence completion task. (we want to know the baseline before we treat)
Historical Research
Collection of data for analysis from existing record Purpose is to construct theories and develop hypotheses based on historical evidence Example: evaluate clinical outcomes
Developmental:Semi-cross sectional
Combine features of both longitudinal and cross-sectional designs Includes 2 or more groups of participants who represent different points on developmental continuum (similar to cross-sectional) However, semi-longitudinal follows participants for an extended time period (3-12 months) Group equivalency is still a threat but less than in longitudinal studies Compromise between cross-sectional and longitudinal studies ***Group equivalence is still a threat here (but less than cross-sectional)
Simulatanous treatment design
Compare effects of alternate treatments on a DV Expose participant to two different treatments at same time Order of treatments are counterbalanced to control for extraneous influences Participants can be exposed to both treatments in same day; baseline data are collected in both sessions prior to introducing treatment ***pg 96 and pg 97 comparing effect of alternate treatments on one behavior. exposing a participant to two different treatments at the same time. you may have more than one participant. still dealing with multiple interference effect.
Time Series Design
Comparisons are made by presenting alternative conditions to a participant at different points in time. Possibly alternating treatment and no treatment conditions across time Look at patterns of data across the different conditions
Qualitative Research Designs: Discourse Analysis Method
Concerned with language use beyond boundaries of utterance or sentence Conversational analysis including strategies with typically developing children and adults and those with disorders ***such as Browns type token ratio and pragmatic analysis.
Correaltional research in CSD
Correlation of .30 is usually meaningful; correlation of .70 or higher is nearly always meaningful The correlation means that the measurement of one variable explains something about the measurement of the other variable Statistical analysis in correlational designs depends on correlation coefficients Pearson product-moment correlation yields a correlation coefficient referred to as "r". Based on continuously measured data To be able to use the Pearson for analysis: measures should be approximately normally distributed, variances of two measures are similar, relationship is linear, sample is representative of population, variables are measured on interval or ratio scale. ***Pearson example r = .2 etc. Data must be either interval or ratio data (to be a Pearson). Must be normally distributed. Linear.
Credibility/Transferability of Qualitative Research
Credibility: believability of results Look at integrity of observations- length, persistence, triangulation of data by multiple sources Referential adequacy: setting aside unanalyzed data and analyzed after conclusions are drawn to see if they match Member checks: sharing conclusions with participants - assessing match between conclusions and participants' reality ***credibility-what has been done to get this information? how long did it take? did they verify what they were researching??
Changing Criterion, Single Case Design
Design using baseline and treatment phases to show treatment effect while controlling extraneous variables Useful for solving clinical problems where skills are practiced and achieved gradually Sequence of events: A-B1-B2-B3; A is baseline and intervention phase B is divided into subphases Treatment subphases have gradually changing criterion; criterion is gradually changed at end of each intervention subphase- typically increased Criterion is changed throughout intervention phase until terminal goal is achieved ***Pg 94. figure 6-3 do expect that patients will go through phases. someone with a tramatic brain injury--initally want them to be quiet for thirty seconds, or sit in their chair and pay attention, once reached you change the period of time they attend to the task (change the criterion a little and that will be the B2 phase and treatment is divided into sub-phases). patient may not be able to reach goal right off the bat.
Qualitative Research Designs: Un-structured In-Depth Interviewing Method
Data gathering with open-ended questions, no formal structure, with iteration of questions, answers, and generating more questions based on previous answers Use of focus groups for broad topics/concepts Still must have preplanned script of specific issues and establish goals for type of information gathered
What is Qualitative Data?
Depends on descriptions, categories, words Observation of events; asking open-ended research questions Recording what is observed, said, done Interpret the data Further observations and asking more questions Development of formal theories to explain the data Formulate conclusions and generate hypotheses Usually detailed, in-depth descriptions of phenomena being studied Data comes from field notes, clinical records, audio/videotape transcriptions, photographs Purpose: describe, explore, or generate ideas, thoughts, feelings and to form hypotheses that can be tested by quantitative research designs Qualitative research focuses on understanding individual differences and similarities
Correlational research
Describing the relationship between two or more variables Then two variables are tested and determined to have a relationship, that are correlated Correlations have direction in that they can be positive or negative Positive: individuals score high (or low) on one variable and similarly on the other variable Negative: individuals score high (or low) on one variable and the inverse (opposite) on the other Correlations also have strength: weak to strong Can range from anywhere between 0 and 1.00 ***One feature of correlations—they have direction have a range anywhere from 0 to 1.00 (a weak correlation .20, a strong correlation .8)
Writing the Research Method
Description of participants: ethnicity, age, gender, intelligence, socioeconomic status, etc. Description of equipment and stimuli used in study Procedures used to collect data: provide this information in a step by step manner with a lot of detail
Address practical significance and Clinical Importance
Efficacy: how well something works in the laboratory or the controlled environment - ideal situation Effectiveness: how well it works in clinical or routine setting Need for confidence intervals: measure of the precision of a statistical estimate - basically, how much stake can you place on the results Effect Sizes: measure of practical significance- looking at extent of change; comparing two different measures and dividing by the variance; degree of overlap refers to amount of variation shared by the two measures LESS overlap means greater difference between two measures (treatments, groups, variables, conditions, baseline vs. treatment, etc.) ***Next step after effectiveness—is it efficient? Someday we should be able to find out how long a specific patient should be in treatment.
Replication Designs: Systematic
Establish generalization of results to other settings or persons How robust is the treatment effect Duplicate a previous study but also incorporate a few variations in the design adding more participants, using different participants, using different settings
Replication Designs: Direct
Establish reliability of previous results Must duplicate study in every detail Only variation is use of new participants Direct replications are typically accomplished by independent teams of researchers to avoid experimenter bias
Writing the Discussion
Evaluate and interpret implications of results Statements of support or non-support for research hypotheses Comment on importance of findings
Multiple Baseline Designs
Examining performance of at least two target behaviors simultaneously. Target behaviors need to chosen carefully: need to be independent of one another Thus change in behavior being directly treated should not yield a change in the untreated behavior Avoids ethical concern regarding withdrawal of treatment because no withdrawal of treatment is needed ***examining performance at at least 2 target bahaviors simultaneously. difficult to choose the behaviors. Hardest part about setting up multi baseline design. treating behavior A and not B, should see a change in behavior A and not B.
Multiple Baseline Design
First, collection of baseline data (A phase) Then treatment is introduced to the first of the targeted variables regardless of behaviors, participants, settings (B phase - target variable 1) During this time, collection of baseline data is continued for the other behaviors, participants, or settings (A phase - target variable 2, or 3 if applicable). Once performance for the first targeted variable reaches some predetermined criterion (e.g. 80% accuracy), treatment is then introduced to the second targeted variable (behaviors, participants, or settings). (B phase - target variable 2)
Ethnography
Focus on entire CULTURE of a people Typically participant observation; researcher is immersed in culture and becomes active participant To gather data, may use: in-depth interviews, focus groups, document analysis, audio/videotape analysis, lamination Use of triangulation: comparing/contrasting data collected at different times, by different methods, different locations Lamination: cross referencing of findings through data collection, forming of tentative conclusions, and then asking participant their thoughts about the interaction No preset limits as to what is observed; no prescribed ending point ***most well know approach. researcher is a participate in the culture.,,immerses oneself in the culture. Traiangulation-used by psychologist,legal. Lamination-cross referencing by data collect from several different sources.
Time Series Design
For external validity or generalization of findings, depend on replication with additional participants Thus less rigorous relative to overall generalization Require decision making regarding the course of the study as it progresses ***you can change stimuli or requirements depending upon the individuals performance. every treatment should be a mini within subject design.
Baselines in Single Case Design
For reliability and validity purposes, minimum of 5 baseline observations are usually recommended However, # of baseline observations depends on: Stability or degree of variability of the dependent variable Time constraints of introducing treatment In baseline, if observed behavior is erratic or there is no stable pattern, would be difficult to interpret your data once treatment is introduced If observed behavior is stable, intervention or treatment phase is then introduced ***taking 2 consecutive baselines from one session to the next can get variability, thats why it is good to get several baselines.
ABA/ABAB Single Case Designs
In ABAB, second B phase is reintroduction of treatment The second treatment phase addresses ethical issue of concluding treatment without benefit of intervention ABAB also replicates the A-B comparison Offers additional support for cause effect relationships between IV (intervention) and DV (what you are measuring) ***we are looking for maintenance of the skill. ABAB reintroduces the treatment. ethics of not completing treatment before intervention people like ABAB because if gives you cause and effect. page 88 figure 6-1 (good example of a single subject ABAB design) you could at long term. you could measure one right after another and then come back 2 months later drawback...the task may be too low level after a long period of time.
ABA/ABAB Single Case Design
In ABAB, treatment is re-introduced in second baseline. If treatment is efficacious, then we expect to see an increase in performance in second B phase ****is reintroduced after second A (after second baseline)
Developmental: Cross-sectional
Includes several different groups of participants with each group representing a different point along the developmental continuum Time period for observation of each group is short Lessens degree of possible participant attrition and demand on resources Limitations: Nonequivalence: because there is inclusion of groups with different chronological ages, participants must be carefully matched to insure equivalence in all respects except age Sometimes weak design and threat to internal validity because there is not ongoing analyses of behavior
Qualitative research designs: study methods
Intensive observation of person, topic, location, event Generates theories and clinical practice ideas Can focus on clinical cases, geographic locations, specific topics or events ****you are not necessarily do an experiment. you are just recording information about this person. Not reporting future events...focusing on the here and now. case study--what is happening now and what is of interest in that individual.
Simultaneous Treatment Design
Intervention phase B begins with first treatment introduced in session one and second treatment introduced in session two If one treatment is superior, this treatment may be continued in both sessions in phase C Intervention is concluded when treatment effect is clearly demonstrated Then optional phase D- maintenance phase to examine carryover effects
What is a survey?
Interviews: more personal and allow interviewer to probe for more information or ask follow-up questions Easier for participants to complete Limitations is that they are time-consuming and interviewers must be trained
Survey research design
Investigate characteristics of population by collecting representative samples Census: each and every member of a population is surveyed Sampling methods: Simple random sampling: every member has an equal chance of being included in sample Identify members of the population Randomly select a sample of members from the population Concerns about bias: failure to identify all members of a population; choosing a sample based on convenience; constituting a sample from volunteers
What is the study/report about?
Is it found in title, key words, abstract, and/or statement of purpose? Is purpose stated in clear and concise words? Abstract: should be brief and comprehensive summary of article; is it: Accurate, self-contained, concise and specific, coherent and readable After examining this information, reader should know if article is relevant to their problem
Phenomenology
More a philosophical understanding of the world than methodological approach Reject speculation and acceptance of unobservable phenomena Oppose naturalism: metaphysical theory that all phenomena can be explained in terms of natural causes and laws Only objects in the world can be known Focus on phenomena as immediately experienced
Editing Stage
More to do with mechanics of writing form and less to do with what is said relative to content Cosmetic improvements in paper: spelling errors, punctuation, grammar Improving writing style: Omit needless words don't comment on your own writing Use active voice rather than passive Use past tense when reporting previous research of others Check to see that all references all listed in reference list
What is a Qualitative research approach?
Naturalistic inquiry: focuses on persons, topics, locations, events in natural setting - classrooms, homes, neighborhoods Inductive analysis: begin with specific examples/facts, formulate questions, ending with general theories Holistic perspective: each action is viewed as part of the whole. Thick description: gathering detailed information about phenomenon of interest, triangulating data from multiple sources Personal contact/insight: awareness of personal biases, acknowledging them, and tempering their influence
Field Research
Observation of phenomena in their natural state or context Collect detailed notes, categorize/analyze in a variety of ways Collecting data in naturalistic settings ***Ethnography fits with field research and and collecting data in natural environments.
Developmental Research: Longitudinal
Observe same participants over extended period of time, typically 1 or more years in length Exact length of study is dependent upon nature of behaviors or traits to be studied and goals established by researchers Initial plan must indicate length of study and the specific intervals for measurements Limitations: Participant attrition - length of study increases risk of losing participants - threat to internal validity Require substantial human resources and finances to sustain study for its duration
Writing/drafting stage
Organize research product Then adopt simple and direct writing style Start strong Tell readers why they should be interested Make sure article or report does what it says it will Make sure literature is focused, complete Explain what your results mean Consider alternative interpretations of the data End strong and state a clear conclusion
Writing the title and abstract
Write after introduction, Method, Results, and Discussion Title should summarize main idea of paper simply Should be concise, name variables, indicated what was studied, mention population studied Sometimes titles ask a question Abstracts: brief, concise and comprehensive summary of research Should specify research hypothesis, purpose, or questions Provide highlights of method Provide highlights of results
Description of participants
Procedure for selecting participants? Procedure for assigning participants to groups? Is control group equivalent to experimental group? Demographic characteristics of participants - age, gender, education, etc.? Participant attrition during course of study? Sample size? Adequate for making conclusions?
Combining Qualitative and Quantitative Methods
Qualitative models serve for first stage of knowledge building to discover key issues and elements for subsequent study Some models conduct qualitative methods (case studies, etc.) in conjunction (parallel) with other methods Within same overall project, methodology can include both quantitative and qualitative approaches ***qualitative--first stage of knowledge. we want to discover key elements. We can do this in conjunction with other methods. most of time we are using qualitative methods as a stepping stone to quantitative methods.
Survey Research Design
Sampling error or bias can be minimized by gathering a sample of adequate size Sampling error is reduced through stratified sampling method divide population into logical divisions to insure that each division or strata is fairly represented in sample
Writing the Appendices and Appendex
References appear in a list at end of research paper Should account for every citation appearing in the body of paper Should use APA Publication Manual style Appendices: Optional but should include if needed critical material that does not fit in the body of the research paper: samples of raw data, stimulus materials, specific directions given to participants,
ABA/ABAB Single Case Design
Remember, introduction of second baseline or A phase tests assumption that first/original baseline was an accurate predictor of future performance without treatment Second A phase: withdrawal phase - tests internal validity of design Rules out effects of history, maturation, etc. We would not want performance to return to baseline levels However, we do expect some decline in performance when treatment is withdrawn ***we dont want performance to return to baseline because we would be out of the job. normal to see a little decline once treatment is removed example: athletic taking a break. when they come back, they wont be as good.
Replication of Single Case Design
Repeat treatment design with other individuals with exactly same conditions Building external validity Synthesis of results from small studies or multiple single case designs builds evidence for effectiveness of a particular treatment Best scientific evidence is still randomized controlled experiments with large representative samples of participants ***large representative difficult because you dont always have access to a large sample.
Writing in Research Design: Overview
Research process involves planning, implementing, and dissemination Dissemination: act of sharing research methods, results, and conclusions with colleagues and others Disseminate research to: Contribute to scientific foundation in CSD Stimulate further research in area of study Test results and conclusions within the community of researchers and consumers Citation rate: number of times research product is referenced by members of the research community Use of research product in SLP and/or Audiology
Qualitative Research Designs: Participant Observation Method
Researcher becomes a participant in culture or context that is being observed Researcher must acclimate to situation and become accepted member in scene Demands much time and resources Useful for observing cultural variables, language development, other social behaviors from an insider's perspective ***useful for observing cultural variables and language development. becoming a participant observer. Have to rely on a lot a different resources. Cant be taking notes while you are participating.
What do the results mean--Discussion
Researchers examine, interpret, qualify and draw inferences from results Need to be consistent with reported results Questions: Are importance of findings discussed and are claims consistent with results Are unexplained results discussed and explained? Are results explained relative to existing theories? Do conclusions match results in results section? Are cautions/limitations of conclusions provided? Are clinical implications discussed if relevant? Are directions for further research provided?
What was found--Results
Researchers typically use tables, figures or other graphical means to enhance clarity of presentation of results; use of confidence intervals Is there clear presentation of results in tables or graphs? Is there verbal description of results relative to what was analyzed? Were correct statistical tests used? Are descriptive statistics provided and appropriate? Are results reported in enough detail to answer research questions? Are the analyses sufficient to address questions of clinical and practical importance?
Evaluating Research in Practice for CSD:::: Information Literacy
Set of abilities requiring individuals to recognize when information is needed and have the ability to locate, evaluate, and use the needed information effectively Information Literacy Standards: Determine nature/extent of information needed Access needed information effectively/efficiently Evaluate information and its sources critically Use information effectively to accomplish specific purpose Understand economic, legal, and social issues surrounding use of information ***Very important. Experience having to write a paper...you research material.
Qualitative Research Designs: Kinesic Analysis Method
Study of what is communicated through nonverbal movements, gestures, postures Usually combined with other qualitative approaches (discourse analysis) to triangulate data Data often arranged in categories: emblems, illustrators, affective displays, regulators, adaptors Cultural constraints ***affective displays-facial expressions regulators-things people do to keep a conversation going or terminate a conversation. Adaptors-things we do ourselves to communicate with others (moving hair out of face, taking off jacket because its warm) cultural constraints- making certain movements or using language in different contexts can be proper/improper. some people just do this type of analysis.
Writing Research Results
Summarize data collected and statistics or data analysis used Include tables and figures to display results Report statistical significance, confidence intervals, effect sizes Explain the results - explain statistical findings, describe the results Description of individual participant performance if appropriate
Multiple Baseline Designs
Target variables/behaviors: Can be Different behaviors for one participant observed in one setting (behaviors) Different participants observed in one setting for one behavior (participants) Different settings with one participant and one behavior (settings) ***Page 93 figure 6-2. most common is the first behavior.
How was the study conducted?
This information is in Method section of an article; typically contains 3 subsections: Description of participants Description of apparatus or equipment and materials/stimuli used in study Description of the procedure - step by step discussion of how the study was carried out
Simple Case Studies
Thorough description of one or more individuals; more qualitative; lack controls of experimental designs; do not have rigorous standards of scientific inquiry; but excellent in generating ideas
Case Study/Ethnographic Designs
Thorough descriptions of one or more participant's traits and behaviors Case study: single participant Describing unusual clinical cases Ethnographic research: targets a group of individuals sharing common bond Cultural traits or specific disability
Writing Process: pre-writing
Time to generate ideas Mentally rehearse organization and writing of research product Adopt APA Publication Manual style Plan for Introduction, Method, Results, Discussion , Abstract, Title, and list of References
Credibility/Transferability of Qualitative Research
Transferability: Extent that results can be transferred from study to other persons, locations, events, contexts Based on consumer's local perspective and external consistency of results and conclusions
Correlational Research cont.
Typically do not imply causality because they do not meet requirements, especially requirement that no other variable can account for the cause However, if strong relationship is found between two variables, special correlational designs can evaluate causality Path analysis: based on a regression model in which one variable can be predicted from another variable Criterion variable is what is being predicted Predictor variable is used to make prediction Multiple regression: extends path analysis of regression and prediction by adding several more variables Factor analysis: involves large number of correlated variables and seeks to identify common underlying factor ***Besides strength and direction, one more feature—significant. (is it statistically significant). Example: .3 was significant, but not strong. (for example a .3 could be significant if there were a lot of participants).
How does the study fit into what is already known?
Typically, this information is found in introductory pages of article - review of literature or introduction or background information This section should: Introduce problem, develop the background, state purpose and rationale Literature review should describe past research results relevant to the target problem Literature review should conclude with statements regarding the RELEVANCY of their study - why is it being conducted? Why is it important?
Replication in CSD
Used to: Establish reliability of results Establish generality of results Establish applicability of results to real-world situations Establish validity and generality of a concept
Time Series Design
Usually focus on single individual continuous observation of a dependent variable with introduction of a specific intervention or treatment Focus is on individual performance Compare different conditions within one participant For internal validity, rely on comparisons between different conditions or phases ***internal validity, individual being compared to themselves. comparing participant at diff points in time treatment and no treatment could equal 2 different treatments. you are comparing treatment to baseline. (next slide)
Efficacy of treatment in Single Case Design
Visual inspection, statistical procedures, both Visual inspection of graphing of data; changes/no changes in DV are readily apparent; should be approximately an overall 60% change in behavior ***Hough-thinks this 60% is too general. if a patient starts out at 20% but now at 60% accuracy and they are consistent. Depends on where you are starting with a patient and how involved. this is a massive generalization.
Evaluating research reports and articles
Was problem addressed through clearly formulated questions? Was qualitative approach appropriate? How were setting and participants chosen Researcher's perspective? Methods used for collecting data - provided in enough detail? Are results credible, clinically important? Conclusions drawn, justified by results?
What do results mean--Discussion
What is a clinically important change relative to the study? Is the benefit of the change cost effective? Is there an alternate treatment that may provide a similar benefit at a lower cost? Is the treatment transferable to my clinical setting? How effective is the treatment likely to be in my clinical setting? What are the economic considerations for my clinical setting? Evidence Grading: use of checklists ***Evidence grading: deciding how well something was written.
Description of Procedure
What wad done, how it was done Needs to presented in enough detail for replication by others or transfer to clinical setting Procedure section should: Summarize each step in carrying out study Describe instructions to participants Describe specific experimental manipulations
Correlational research in CSD
When measurements are in ordinal scales, such as ratings, use the correlational procedure of the Spearman rank correlation Index of determination: r2 Arithmetic square of the correlation coefficient Proportion of variance in one measure that is attributable to the other measure: degree of variance that is shared by the two variables The more that two variables have in common, the better the prediction of one variable from the measurement of the other
Time Series Design
also referred to as within subject designs or single case designs include controls to meet standards of scientific inquiry observe series of behaviors across some period of time permit cause-effect conclusions Conclusions are less robust than randomized controlled group designs - thus, considered quasi-experimental designs Used to gather preliminary data before undertaking randomized controlled group designs Provide strong evidence for clinical practice ***somewhat longitudinal cause-effect (Im going to this therapy and this is the effect that Im getting. still considered Quasi-experimental. you are going to do a baseline measure.
