Research Roles, Sampling, & Sites

maximum variation sampling

seeking individuals who have extremely diferent experiences of the phenomenon being studied

homogeneous selection

seek informants who have the same experience, experienced the same phenomenon

theory-based selection

researcher selects only individuals who exemplify a partcular theoretcal construct for the purpose of expanding the current understanding of a theory (e.g., constructs of co-occupaton)

confirming or disconfirming cases 

researcher looks purposefully for the informant who will support or challenge an emerging interpretaton

determining sample size

‒ No standards for the ‘right’ number

‒ Focus on quality of data

• Depth vs breadth

gaining access

entry point into a qualitatve inquiry & afects the selecton of subjects

‒ “Gatekeeper”

domain analysis 

the critcal process of selectng & adding pieces of informaton through interview & observaton & analyzing it for further discovery

ethical responsibility

awareness of need to engage in research for practce development, innovatons (appropriate care)

professional responsibility

obligaton to use research to enhance the quality of decision-making

research producers 

Academics, practitioners, students

Training: doctorate, postdoctoral training

Activities: 

• identifity original research questins and design methods for answering them 

• secure funding research 

• supervise/oversee the implementation of research 

• work with interdisciplinary collaborators, statisitcial consultants, and other with specific research expertise 

• prepare research reports for presentation or publication 

research collaborators 

Supportve role to conduct of research, may include intellectual contributon

‒ Effective: negotaton, planning, good communicaton, trustng relatonships

Training: professional training; specialized training related to study involvement 

Activities: 

• serve as subjects/participants 

• refer and screen potiential subjects 

• collect data 

• implement services that are being tested

• provide clinical advice/expertise 

• help negotiate the politics and administrative process 

• help identiffy a research question and interpret results 

research consumers

Utlizes research

Training: professional training training in critical appraisal skills

Activities:

• read research and use finidings to guide practice 

• identify practice knowledge gaps 

• complete and present/publish critical appraisals 

research advocates

Provide momentum & support for research

‒ Local, state, natonal, internatonal

Training: appreciation fo the value of research that can accrue from traiining and/or experience 

Activities: 

• serve in professional associations that fund/support research 

• lobby policymakers and legislators 

• ask employers for resources 

• donate moneya nd encourage/support colleagues involves in research 

• comment on research priorities or goals 

resistance to research

• Lack of exposure to research

• Concerns about lack of tme for research

• Anxiety about the perceived or actual lack of skills or knowledge to conduct research

• Clinical practce guided by personal frames of reference &/or worldviews outside professional approaches

• Choice not to engage in research

research subjects

groups that represent a larger populaton with a common phenomenon under study

‒ Sample- Specifc, clearly writen

inclusion criteria 

the traits that the researcher has identfied as characterizing the populaton

exclusion criteria

the characteristcs that will prohibit the subject from being an appropriate candidate for the study

subject pool

those identified as eligible to participate

defining the sample

Does the study research queston infuence the subject pool? PICO Review!

And the research approach: (quanttatve, qualitatve, efcacy,...)

‒ Access to populaton

‒ Resources

• Ensuring representaton as per funding requirements

‒ Race, sex, ethnicity,...

steps in sampling 

1. Define the populaton of interest

• Literature review

2. Consider the unit of analysis

• Typically the subject, can be setngs

3. Identify a performance site (clinic, school, SNF, etc.)

4. Develop a sampling approach

• How researcher will select/recriuit the sample

• Power analysis, data saturaton (QUAL)

5. Implement the sampling procedures

sampling in quantitiative research 

• Determined before a study begins

• Defne target populaton –a the subset s the sample/partcipants

‒ Accessible populaton

‒ Generalizability

• Need to be able to accurately make conclusions about the populaton bystudying the sample

• The specifcaton of the number of persons in the sample is established afer thedesign is determined

external validity

the extent to which we may generalize a research fnding to different person, setngs, or tmes

generalizability

representatie sample increases _

sampling bias (selection bias) 

criteria used to recruit & enroll subjects into study/study conditons that are inherently diferent

‒ Overrepresentation

‒ Underrepresentation

major threats to external validity

‒ People (some more invested in results than others)

‒ Place (SES)

‒ Time (pre- vs post-COVID)

probablity

random sampling

randomization 

‒ considered the cornerstone of quantitatve research

Balances measured & unmeasured characteristcs that affect the outcomes of a study, allows for masking, & provides a basis for inference

• Method to address selecton bias

startified sampling

selecton from identfed subgroups of a populaton

• searchers divide a populaton into homogeneous subpopulatons

called strata (the plural of stratum) based on specifc characteristcs (e.g., race, gender identty,

locaton, etc.) every member of the populaton studied should be in exactly one stratum

systematic sampling

equivalent to random sampling as long as there is not reoccurring patern or order in the listng

cluster sampling 

groups or programs, rather than individuals, are selected, & every member of that group or program is invited to partcipate in the study

“Cluster” may already exist, e.g., schools, SNF, hospital, etc

nonprobability 

non-random sampling

‒ Clearly defne the process of the sampling

‒ Acknowledge the limitatons of the sampling procedure

‒ Justfy why the sampling limitatons do not jeopardize the research queston being answered

purposive

the deliberate selection of individuals by the researcher based on certain predetermined criteria (e.g., recipients of Go Baby Go! Cars –a consider inclusion criteria)

convenience

use of volunteers or easily available subjects such as a group of studentsin a program or clientele in a clinic

snowball or network 

initally identfed subjects provided names of others who may meet the study criteria (homeschool handwritng)

quota

used when diferent proportonsof subject types are needed so that there is appropriate representaton in the sample that may not be atainable with purposive or convenience sampling

crowdsource or panels

online sampling

sampling error 

the diference between the values obtained by the sample & the actual values that exist in the populaton (likely do not have this informaton so may estmate).

random error

an error that may occur by chance when administering an assessment (study of driving eval norms and a Formula 1 driver –a outlier)

systematic error

onsistent or predictable errors; they occur when an assessment misestimates the true score by a consistent amount & in the same directon (general educaton school classroom vs children receiving OT services for handwritng clinic in the evenings)

‒ Sources

• Volunteers (diferent from those who do not)

• Groups that are available & convenient

correlational research 

conventon of minimum 30 subjects

survey research

conventon of minimum of 100 subjects in each major subgroup & 20-50 in each minor subgroup whose responses will be analyzed

casual comparative

minimum 15 subjects per group

statistical power 

likelihood of finding a significant difference between groups or association between variables when one exists

‒ Directly related to sample size

‒ 4 components: sample size, efect size, alpha level or level of significance, power of the odds of observing a treatment effect when it occurs

effect size

defnes the extent of diference between two means or the size of the relatonship between variables in a research study

‒ Smaller efect size, need more subjects

level of significance

typically .05: means researchers are willing to accept a 5% chance that they will fnd an efect by chance when there really is no true efect

type I error 

mistakenly accept the alternatve hypothesis when the null hypothesis is in fact correct & should be accepted (in reality, there is no diference between groups s false positve)

calculating sample size

use expected efect size, alpha level, & power

‒ As the sample size increases, the power increases

‒ If variaton in outcome decreases, the power increases

‒ If variaton in outcome increases, the sample size needs to increase

‒ The power increases as the efect size increases

‒ If the efect size decreases, the sample size needs to increase

facotrs affecting sample size

‒ Subgroup analysis (e.g., by gender, hand dominance, age)

‒ Expected atriton (dropouts)

‒ Reliability of measures

apporpriateness 

the identfcaton of partcipants who will best inform the researcher about the phenomena under inquiry

adequacy

enough data will be available to provide a rich descripton of the phenomena of interest

Goal is saturaton: afer contnued interviewing &/or observaton, no additonal informaton is gained

unit of analysis

people-focused or structure-focused (e.g., programs, organizatons)

‒ Culture

‒ Geographic or organizatonal locaton

‒ Time or event-related experience (e.g., 9/11)

‒ Personal experience of a unique conditon