cdis 5400 dr brenda louw 2010 research methods in communication disorders
TRANSCRIPT
CDIS 5400Dr Brenda Louw 2010
Research Methods inCommunication Disorders
Objectives
Demonstrate: an awareness of the researcher-clinician
relationship knowledge and understanding of scientific
methods and how they are applied in research Understanding of research ethics formulate answerable research questions
Overview
Introductions Course overview Syllabus
Research-Practice relationships
Method of science Research ethics Research
questions & hypotheses
Research topics
Introductions
Personal research experiences
or Course
Expectations
Instructor research experience: Teaching Doing Learning
Research goals at graduate level Help students become competent
consumers of literature Help students become competent users
of research technology for clinical decisions
(Bain,1991)
Competent Consumers of research
Critically evaluate the literature Determine cautions in interpreting the
literature Determine applications to clients served Determine how procedures might be
modified to better serve clients
Research-Practice relationships in SLP Traditionally : practice informs science
and science informs practice Advances in diagnostic and treatment
protocols for a specific communication disorder derived from scholarly research
Gap between clinicians and researchers: Clinicians :I don’t read journals Researchers: I don’t care if clinicians don’t
read my papers (Logeman,2000) Evidence –Based practice
EBP is an integration of:
(McCauley, 2004)(McCauley, 2004)
Current Best Evidence
Clinical Expertise Client Values, Culture, Environment, Preferences
“When scientific clinicians approach clinical problems in a scientific manner, they are conducting research of the most important type, with the result being the intent of delivering the best clinical management possible.” (Ringel, 1972)
Research Practice relationships Shared interest communication: interests
of researchers and clinicians overlap Research driven communication: research
findings available to clinicians-books, journals, conventions
Practice –driven communication: clinicians express their interests to researchers resulting in initiation of projects
Knowledge acquisition
Method of tenacity: people hold firmly to certain facts assuming those are true
Method of authority: people accept knowledge from an individual, group
Priori method: relies on use of pure reason based on prior assumptions considered to be evident
Method of science: systematic, controlled, empirical, amoral, public, and critical investigation of natural phenomena, guided by theory and hypotheses about the presumed relations among such phenomena.
Schiavetti et al.,2011
Empirical research
Collection of new information or data through observation and measurement of behavior and /or physical properties
In SLP e.g. : speech samples, listener ratings, surveys, test scores, measures of speaker sound pressure levels, tongue strength
The Scientific Method (Williams,2008)
Scientific research directed toward “basic research” or “applied research”
Scientific Research
Basic Research Applied Research- development of a - used to solve some
knowledge base some social or economic
problem
Types of ResearchHistorical Descriptive ExperimentalUses documentary Examines group Examines causa-
evidence to identifydifferences, trends tion through
patterns or relationships observation of
among variables the effects of the
using laboratory manipulation of
measurements certain variables
on other variables under
controlled conditions
Experimental Research
Used to investigate cause-and-effect relationships among variables
Researcher manipulates what subjects will experience
Characteristics of experimental research: starts with a purpose, question, or hypothesis CONTROLS events and therefore observes
CHANGES in behavior is repeatable can measure effects on certain variables by
manipulating other variables must be cautious to control “extraneous
variables”
Descriptive Research
Assessed nature of existing conditions Examines group differences,
developmental trends or relationships among factors
Research uses objective measurements –tests,surveys,naturalistic observations
Does not lead to a cause-and-effect identification between IV and DV (all you are doing is describing the behavior)
continued
Types of descriptive research: comparative research developmental research correlational research survey research case study retrospective research
The conduct of scientific research
Stating the problem Formulating research
question,hypotheis Developing the research method Analyzing results of investigation Interpreting results
Meline, 2010;Schiavetti et al., 2011
Variables
Measurable quantities that may vary or change under different circumstances rather than remain constant
Independent variables: conditions that the cause changes in behavior(conditions researcher is interested in studying)
Dependent variables: the behavior that is changed(observations, measurements obtained)
Eg dleayed Auditory feedback IV) may cause change in speech rate(DV)
RESEARCH ETHICS
ASHA Principles of Ethics I-IV ASHA Guidelines for the responsible
conduct of research: Ethics and the publication process
Ethics principles in research : respect/autonomy Beneficence –act in best interest of participants
Justice- fairnessLevine,2003; Belmont Report
Researcher’s ethical responsibilities
Protection human subjects (Nelson, 2009) Assigning credit for intellectual effort Accurate attribution of ideas and citing sources Accuracy and honesty in reporting information Avoiding conflicts of interest Confidentiality and privacy Evidence-Based practice (Meline,2010) Clinicians adopt scientific method Systematic use of best evidence(high quality
research) to solve clinical problems
Ethical norms/rules derived from principles1 Respect for persons:
Informed consent Privacy, confidentiality, anonymity Withdrawal from participation Honesty Publication and authorship Debriefing Inducement of participation Levine, 2003
continued
2 Beneficence: Good research design Competent researchers Favorable risk-benefit balance Relevance of research Safe guards for vulnerable populationsLevine, 2003
continued
3 Distributive Justice Inclusion and exclusion criteria for
participants Liability of researchers Fair distribution of benefitsLevine, 2003
Institutional Review Boards(IRB) Provide institutional oversight Evaluate proposed studies before
researchers begin collecting data Protecting rights and safety of research
subjects Conduct risk/benefit analysis
Clinical and research questions Clinical questions :
Assessment treatment
Research questions Require answers
Clinical Decisions: Assessment
1. Who is disordered (is there a problem? If so, what is it?) Expresive? pragmatic?
2. What is the etiology? What are the facters contributing to the child having the problem
3. What is the prognosis? What are the childs possiblity of improving and on what level?
Clinical Decisions: Treatment
1. Who should receive what kind of intervention?
2. What should be treated?3. How should we provide intervention?
Alternating treatments design (ATD)
4. Is intervention effective? Single-subject designs
5. When should intervention be terminated?
ABA design
RESEARCH QUESTIONS AND HYPOTHESES Research problem is identified and
defined Options:
Research questions Formal hypotheses Statement of purpose
Asking Answerable Questions
“One cannot get a clear answer to a vague question. The language of science is particularly distinguished by the fact it centers around well-stated questions. If there is one part of a scientific experiment that is more important than any other part, it is the framing of the question that the experiment is to answer. If it is stated vaguely, no experiment can answer it precisely. If the question is stated precisely, the means of answering it are clearly indicated. The specific observations needed, and the conditions under which they are to be made, are implied in the question itself.”
Wendell Johnson (1946)
Two Parts to Asking Answerable Questions How can I generate relevant questions
for research? “So what” or “who cares” -- need to
generate relevant questions for which there is a need to answer
How can I structure the research question to make it “answerable”?
Part I: How do I generate questions for research?
We use the scientific method as a set of rules for asking and answering questions the scientific method suggests rules or
criteria for formulating questions, making observations to answer them, and relating the answers to the existing body of knowledge
questions formulated for research should be unambiguous, clearly defined, and precise
How do I decide what I want to investigate?
FIRST, must decide what to investigate what question do you want to try to
answer
deciding what to study is important because it affects the impact of the research as well as the likelihood that you will complete the research
How do I decide what I want to investigate?
Impact - the effect that answering a particular question will likely have on the clinical process in SLP or AUD
Likelihood of completing the research - several factorsseveral factors must be considered:
1. The time required to make the necessary observations -- the longer you take to make the observations, the less likely you are to complete your research
Continued..
2. The complexity of the procedures required to make the necessary observations
the more complex the procedures, the less likely you may be in completing the research
3. The availability of subjects if subjects aren’t available in sufficient
numbers within a reasonable time period, the probability is reduced that the project will be completed
Continued…
4. Time available for the project If your available time is limited,
then the possibility of extending your graduation exists
5. Availability of supporting personnel or equipment statistician technician
Continued…
6. Available funding if the costs for completing the
project are high, the less likely it will be completed
7. Personal interests and competencies identify an area within CDIS that is of
most interest to you and one in which you have the training and expertise to be able to carry out the project
Determining a Research Area Use personal interests
examine your clinical experience in the area of interest for questions that seem worth answering and then search the relevant literature to determine whether they have been answered ask yourself, “what clinically relevant questions
have I wanted answers for but there were not data to answer it?” What questions am I assuming have been answered, but haven’t?
Narrowing Down a Research Question
Ask questions that are based on answers of authority or tradition rather than empirical data.
EX. How do we know that a procedure does what it is supposed to do?
For example:
One might ask whether the number of sessions per week of a particular length is optimal for reducing the severity of their clients’ communicative disorder. Perhaps the sessions are 1) longer and more frequent than necessary; 2) longer but not more frequent than necessary; 3) less frequent but no longer than necessary; 4) shorter and less frequent than necessary
What approaches can you use to generate a question? Review literature in the area
selected in order to identify: questions that have been formulated but
only partially answered questions that have been formulated but
not answered questions that have not been formulated
but should be
Formulated questions that have been partially answered Maybe didn’t use a random sample of
subjects to answer the question
observations may not have been reliable, valid, or both
experimenter bias
Formulated questions that have not been answered
May be found in discussion section of articles review papers in an area of interest which
summarizes and integrates the research in a particular area
best way to locate review papers is through an abstracts journal (psychological abstracts, dsh abstracts, language and language behavior abstracts, index medicus, and dissertation abstracts publications in which individuals or committees
indicate what research they feel is needed in particular areas
Questions that have not been formulated but should be The strategy here is to:
1. Do a careful search of the literature in the area to determine what is known search can be facilitated by using abstracts
journals organize the literature so that it is apparent
what is known about the area 2. Formulate questions which, if answered,
would be likely to increase our understanding of the area.
Now that you’ve identified an area …
Generate one or more questions to answer that are relevant to the area these questions define the research topic they indicate what the investigator is
trying to find out you can usually determine from the
question what the possible answers could be
questions specify what has to be done to answer them
Part II: How can I structure the question to make it answerable?
You need to structure a question so it is obvious what observations have to be made to answer it EX: 1. Is hypnosis effective in treating
stuttering? 2. Is the post-hypnotic
suggestion, “you will not stutter any more”, effective in reducing stuttering frequency?
you must have observations that are possible to make
Asking an Answerable Question
The question must be clear and unambiguous EX: Do persons with cleft palates have
normal intelligence? What segment of the CP population are you
investigating (lip only, hard palate only, soft palate only, hard and soft palate, etc.)
what is meant by intelligence (performance on what task or test?)
More answerable question: Are five year old children who have clefts of the palate only more likely to have below normal intelligence as measured by the Stanford-Binet than five year old children who do not have this condition?
Asking Answerable QuestionsLess Answerable
Ques Does drinking
alcoholic beverages reduce stuttering?
More Answerable Ques
Does drinking 5 highballs in a period of 2 hrs reduce the stuttering frequencies of most adult stutterers?
Asking Answerable QuestionsLess Answerable
Ques Are home
assignments helpful for correcting articulation errors?
More Answerable Ques
Are homework drill sheets for final consonant deletion helpful in the transfer stage of therapy in correcting this error?
In Sum ...
To ask answerable questions, must specify: WHO subjects are WHAT conditions are WHAT measurements will be
Hypotheses
Formal statement of predicted outcome Null hypothesis: H0
Stated in negative Assumption that results will yield no significant
differences between groups, variables reject or fail to reject hypothesis H0 There is no relationship between CV ratio and
MLU in preschool children from ages18-36 mos Research question : What is the relationship
between CV ratio and MLU in preschool children from ages18-36 mos
Continued:
Research hypothesis: H1
Alternate hypothesis Stated in positive Can state direction H1There is a significant relationship
between CV ratio and MLU in preschool children from ages18-36 mos
H1 Directional There is a significant positive relationship between CV ratio and MLU in preschool children from ages18-36 mos
SELECTION RESEARCH TOPICS See handout for topic summaries Consider your interests, co-investigators Consult faculty September 29 time line, project
schedule
Reading assignments
ASHA Code of Ethics (http://www.asha.org/docs/html/ET2003-00166.html)
Schiavetti et.al .2011, Chapter 1, pp-24; Chapter 5, pp173-175
Nelson, 2009, Chapter 2, pp11-29; Chapter 3, pp31-49.