CDIS 5400 Dr Brenda Louw 2010. Terminology Populations and Samples ◦ Sample characteristics ◦ Selection criteria ◦ Sampling methods ◦ Sample size.

Research ParticipantsCDIS 5400

Dr Brenda Louw 2010

Terminology

Populations and Samples◦ Sample characteristics◦ Selection criteria◦ Sampling methods◦ Sample size ◦ Power analysis

Research ethics

Overview

Schiavetti et al. 2011, Chapter 5,pp158-175, Chapter 8,pp 370-373

Readings

Subjects:◦ Traditional quantitative scientific research

Participants◦ Qualitative research ◦ APA 2001 recommendation

Complicating factors◦ Research design terminology e.g. Single subject

design, within-subject design◦ Individuals who participate, but cannot give

Informed consent`

Terminology

Currently :◦ Subjects in quantitative research ◦ Participants in qualitative research ◦ Respondents in survey studies(questionnaires)◦ Examinees in test-developing research◦ Also observers, listeners, judges, raters depending

on research task, design

Gender vs biological sex◦ APA :

gender refers to role, cultural biological sex: female or male

Terminology

Other:◦ Animals◦ Specimens and tissue samples◦ Descriptions required◦ Research ethics and IRB

Conclusion:◦ Irrespective of terminology : describe individuals

clearly, concisely

Terminology

1st step in research design and planning - identify population of interest/ target population

Use of entire population - census Extract sample of target population

◦ make inferences about the population based on data from a sample

◦ Indirect conclusions

Populations and Samples

Sample needs to represent population Representative sample

◦ good match of characteristics of the population◦ Include individuals from each constituency of the target

population –including women and children Sample depends on purpose and methodology Bias in selecting sample

◦ Unbiased sample-all members of population have equal chance of being included

◦ Sources of bias: failing to identify all members of a population using sample of convenience volunteerism

Sample characteristics

De Kroon et al.,2002 LSHSS Investigated language impaired children’s

social pretend play and conversational behaviors

Selected 7 boys as participants-3 LI and 4 normal language development

Is the sample representative of the population of children with LI ?

e.g. Meline 2010

Erler & Garstecki @002) AJA Do women vary by age in the degree of

stigma they attach to hearing loss and HA use?

191 women as participants Do the participants match the stated

purpose of the research?

e.g. Meline 2010.

Establishing selection criteria critical aspect in planning research

Selection criteria important to validity and generalization, transferability of results◦ e.g results due to IV or unreported subject

variables Criteria crucial in between-subject designs-

comparison of 2 groups

Selection criteria

Need to describe, define and defend criteria

Samples not all homogenous and lack of detailed descriptions implies that they are

Sufficient detail required to:◦ replicate the study◦ generalize conclusions to real-life situations

Critical reviews◦ Schmitt & Meline(1990) examined 92 research reports with LI

children-only 30% included information about individual participants

Selection criteria continued

Reilly and Moore,2009 Infant respiratory movements during vocalizations (JSHR, 52,223-239)

40 infants: born within 2 weeks of their due date, weighed at least 6 lbs at bith, had negative medical histories for respiratory, cardiovascular and neurological disease, monolingual English homes.

Exclusion criteria: ear infections in week prior to testing, family history hereditary HL, family members with speech, language disorders

e.g.

Simple random sampling Stratified random sampling Cluster sampling Purposive sampling

Sampling methods

◦ Enhances external validity◦ Improves generalization to the target population◦ Every member of the population has an equal

chance of being selected for the study◦ Use random numbers e.g. assign all participants a

number, consult table of random numbers or random number function in spreadsheets

◦ Time consuming method◦ Seldomly used

Simple random sampling

Divide accessible population into subgroups/strata

Then draw subjects randomly E.g. stratify according to age, sex, socio-

economic level, parent education levels etc.,

and then determine groups

Stratified random sampling

Used for large numbers of participants Start with obtaining random sample of

predetermined groups e.g. classroom , community, clinic

Multistage sampling-e.g. choose school district randomly, then select one school from district randomly ,then select one classroom ,then select the children

Cluster sampling

Research question requires input from special persons, groups

Purposeful effort required to identify them Recruit subjects

Purposive sampling

Broad range of sample size due to diversity of research in Communication Disorders

The purpose of the study, nature of the research design and variability of the data determines the adequacy of the sample size

E.g. ◦ within –subject designs-many repeated

observations and data points-require small sample◦ between subject designs require larger sample size◦ survey research requires large numbers;◦ single subject designs-small numbers

Sample size

Recruiting large samples problematic in communication disorder research

No absolute numbers but variable numbers determined by how large the population is ,how variable the population is for the characteristics under study etc.

Sample size

Sample size contributes to power of the research design

Power is the ability of a research study to detect significant treatment effects when they are present (Meline, 2010).

2 general uses of power analysis◦ a priori –to determine subject sample size to reach a

given alpha level◦ post hoc- to further evaluate research that has

been completed to determine if failure to reject a null hypothesis was related to an insufficiently large sample size

Power of research design

When all else is equal, research designs with more participants will have more power to identify treatment effects

80% power recommended for detecting moderate-sized effects

Computer software to perform power or sample size calculations e.g. STAT POWER, NQUERY

Continued…

Schiavetti et al.,2011 Chapter 5, p228 8 evaluation items in the methods section

Evaluation of subjects/participants when reviewing research

Protection of :◦ human subjects◦ animals ◦ Specimens and tissue samples

Research ethics

Study the methods section of a research article Identify the

◦ size of the sample◦ the method of selection◦ Selection criteria

What is the purpose for including each criterion Is sample size adequate for the purpose of the

study? Is an adequate description provided of the

participants Is there evidence of adequate protection of

Ss/participants

Excercise

Review an article with two or more groups of participants and describe the research design. Why did the researchers include more than one group of participants?

Continued…