Sampling Methodology

1
Sampling Methodology
Quantitative Research Methods MPH, AUA
Spring 2007

• Group 3
• Armen Torchyan
• Stella Arakelyan
• Ruzanna Grigoryan
• Varuzhan Ghazaryan

2
Outline

• Research Questions and hypothesis
• Target/study population
• Sampling frame
• Sample size
• Sample design
• Protocol for interviewers

3
Research Question ?1 and hypothesis

• RQ1 What are the knowledge level of district
  pediatricians/family physicians with pediatric
  background (DP/FPP) with regard to indoor risk
  factors for asthma development?
• H1.1 Percentage of respondents who mentioned
  environmental tobacco smoke (ETS) as a risk
  factor will be less than 75 .
• H1.2 By open-ended checklist more than 75 of
  respondents will report 5 or more appropriate
  correct responses.

4
Research Question ?2 and hypothesis

• RQ2 What are the DP/FPP experiences with asthma
  diagnosis, treatment and referral?

5
Target/study population

• District pediatricians/family physicians with
  pediatric background currently working in
  polyclinics of Yerevan
• Sampling unit - A district pediatrician/family
  physician with pediatric background currently
  working in polyclinic of Yerevan

6
Inclusion criteria

• Those providers who are
• currently practicing as district pediatricians in
  polyclinics of Yerevan
• currently practicing as family physicians with
  background in pediatrics

7
Exclusion criteria

• Those pediatricians who
• have any narrow specialization

8
Sampling frame

• Lists of those polyclinics that provide pediatric
  health care in Yerevan

9
Sample size

• It is calculated on the basis of
• nZ21-?/2P(1-P)/d2 1.9620.50.5/0.092
• n 119
• Z confidence interval (95)
• a type I error (0.05)
• P assumed proportion of providers
• who report about ETS as a risk factor for
  asthma development (0.5)
• d desired precision (9)
• Power 0.8

10
Adjustment of sample size

• Response rate - 90
• Proportion of eligibles - 95
• -absence of a unit for the given period for
  data collection
• -inaccuracy of sampling frame

11
Calculation of adjusted sample size

• n 119 / (0.9 0.95) 140

12
Final sample size

n 140 (10 per interviewer)
13
Sample strategy

• Combined sample design

14
Combined sample design

• Cluster - a polyclinic providing peditatric care
  in Yerevan
• Number of clusters 22 (all polyclinics with
  pediatric care)
• Total number of units in each cluster
• Number of units to be selected from each cluster
  sampling with probability proportionate to size
  of cluster
• Selection of units from each cluster
  convenience sampling

15
(No Transcript)
16
Advantages and disadvantages for recommended
strategy

• Advantages
• Feasible for a given period of time
• Includes all polyclinics
• Easy to compare data between clusters

• Disadvantages
• Participation of a unit is affected by decision
  of authority
• Homogeneity of units within clusters
• Potential inaccuracy of sample frame
• Potential bias due to convenient sampling

17
Protocol for interviewers

• Receive number of interviews, addresses of
  polyclinics to be visited and support letter from
  executive group
• Make an appointment with chief of polyclinic
  beforehand
• Get permission from chief of polyclinic to
  conduct interviews with DP/FPP
• reminding about permission of municipality
• introduce support letter of AUA
• In case of refusal go to the next polyclinic

18
Protocol for interviewers

• In case of permission find available DPs and
  FPPs.
• Skip those who have narrow specialization
• Skip those who have participated in recent
  qualitative research conducted by AUA students

19
Protocol for interviewers

• If number of available providers is what you need
  conduct interviews with all of them
• If number of available providers is less than
  what you need wait for the next shift
• If number of available providers is more than
  what you need do convenient sampling to select
  your respondents (conduct interview in each
  polyclinic during one day, do not return)

20
Protocol for interviewers

• If the eligible provider refuses to participate
  drop her/him out, mention it in the questionnaire
  and go to the next
• In case an eligible provider agrees to
  participate conduct the interview

21

• Good luck!!!