Top Banner
1 Audio Information: Dial 1-855-767-1051 Conference ID 61304911
26

Statistics in Evidence Based Medicine II · Statistics in Evidence Based Medicine II ... Stages of survey research ... commonly used as measures of test-retest

May 04, 2018

Download

Documents

duongthuan
Welcome message from author
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Page 1: Statistics in Evidence Based Medicine II · Statistics in Evidence Based Medicine II ... Stages of survey research ... commonly used as measures of test-retest

1

RizwanaRehmanvagov

Statistics in Evidence Based Medicine

II Lecture 2 Linear Correlation and Regression

Audio Information Dial 1-855-767-1051 Conference ID 61304911

Text Book

Basic and Clinical Biostatistics (2004)

Beth Dawson Robert G Trapp

httpwwwaccessmedicinecomresourceTOCaspxresourceID=62

For more information program materials and to complete evaluation for CME credit visit

wwwepilepsyvagovStatistics

2

Audio Information Dial 1-855-767-1051 Conference ID 61304911

Overview

Stages of survey research

Types of surveys

Questionnaire bias

Measurement data types

Reliability amp validity

Sampling and sample size

Analysis of survey data

3

Stages of the Survey Research Process

4

1 Identifying the focus of the study and method of research

2 Determining the research schedule and budget

3 Establishing an information base

4 Determining the sampling frame

5 Determining the sample size and sample selection procedures

Designing and Conducing Survey Research a Comprehensive Guide Luis M Rea amp Richard A Parker

Stages of the Survey Research Process

6 Designing the survey instrument

7 Pretesting the survey instrument

8 Selecting and training interviewers

9 Implementing the survey

10 Coding the completed questionnaires and computerizing the data

11 Analyzing the data and preparing the final report

5 Designing and Conducing Survey Research a Comprehensive Guide Luis M Rea amp Richard A Parker

Types of Sample Survey Research

Mail out surveys

Web based surveys

The telephone survey

In-person interviews

Intercept surveys

6

Open-ended vs Closed Questions

7

Basic and Clinical Biostatistics Chapter 11

Questionnaire Bias

Effect of varying question wording

Different interpretation of respondents from the questioner

Respondents donrsquot understand the question

Setting in which question is asked

8

Levels of Measurement

9

Nominal

Ordinal

Likert scales

Interval

Likert Scale

Likert scales use fixed choice response formats

Designed to measure attitudes or opinion

A Likert type scale assumes that the strengthintensity of experience is linear

Most often five to seven choices including a neutral choice are offered

Common parametric tests are generally used

but analysis should depend upon the research question

10

Reliability of a Survey

11

Reliability is a statistical measure of how

reproducible the survey instrument data

are

Two survey instruments with same outcome

of interest may have different reliabilities

Types of Reliabilities

Test retest

Alternate form

Internal consistency

Intra-observer

Inter-observer

12

Test-retest Reliability

Test-retest reliability is measured by having the same respondents complete a survey at two different points in time

Test-retest reliability is measured by intra-class correlation coefficient

13

22

2

Bewteenwithin

Bewteen

SS

SICC

Swithin and SBetween are the within and between subject standard deviations

Example of Test-retest Reliability

Juniper et al (1996) evaluated the pediatric Asthma Quality of Life Questionnaire (PAQLQ) by examining reliability in children aged 7 to 17 who had stable asthma

14

Item Within subject SD

Between subject SD

ICC

Overall QoL 017 073 095

Symptoms 022 084 093

Activities 042 096 084

Emotions 023 064 089

Medical Statistics A Text Book for the Health Sciences Michael Campbell David Machin amp Stephen Walters

Internal Consistency Reliability

A set of questions measure different aspects of one concept

Cronbachrsquos coefficient alpha (αChronbach) measures internal consistency reliability among a group of items to form a single scale

15

2

2

11 T

i

Chronbachs

s

k

k

k=number of items si=standard deviation of the ith item sT = standard deviation of the sum score T of all the items

Example of Internal Consistency

McKinley at al (1997) used a questionnaire to measure patient satisfaction with out-of ndashhours calls made to general practitioners They measured aspects such as satisfaction management and the doctorrsquos attitude They found values of αChronbach for each score ranging from 061 to 088 They concluded their questionnaire had satisfactory internal consistency

16

Medical Statistics A Text Book for the Health Sciences Michael Campbell David Machin amp Stephen Walters

Inter-observer(Inter-rater) Reliability

Inter-observer reliability provides a measure of how well two or more evaluators agree in their assessment of a variable A survey can be compared with a gold standard by using Cohenrsquos Kappa (k)

17

Chance

ChanceAgree

P

PP

1

PAgree= Estimated Probability the observerrsquos agree PChance= Proportion expected to show agreement by chance

Example of Inter-observer Reliability

18

Clamp and Kendrick (1998) used a telephone survey asking about safety to 165 families with children under age 5 years who had been involved in randomized trial concerning safety advice They chose a random sample of 20 families from the survey who then received a home visit 2 weeks later to measure the consistency of the response to the questions posed The investigators found that for most questions a high kappa value ( gt059) was obtained

Medical Statistics A Text Book for the Health Sciences Michael Campbell David Machin amp Stephen Walters

Validity of a Survey

A survey is valid when it measures what it is supposed to measure

Content

Face

Criterion

Construct

19

Use of Correlation Coefficient for Reliability and Validity

Pearson correlation coefficients r are commonly used as measures of test-retest reliability inter-observer reliability and concurrent validity

Results could be misleading due to systematic biases

20

Determining the Sample Size

Survey should have enough power to find an effect if there exists one

Q What sample size would be enough to generalize results

Ans Determination of sample size depends upon study outcomes

Researcher must consider

Level of confidence

Margin of error 21

Survey Sampling Methods

Simple random sampling

Systematic random sampling

Stratified sampling

Cluster sampling

Convenience sampling

22

Analyzing Survey Data Descriptive Statistics

Measure of central tendency

Measure of dispersion

Inferential Statistics

Parametric amp non-parametric tests

Chi square

ANOVA

Regression amp correlation

More advanced techniques

23

Openstat A free software

Download

httpwwwstatprograms4ucomOpenStatMainhtm

Tutorial

httpmmc2geofisicaunammxcursosgeoestSoftwareOpenStatATextBookpdf

Youtube tutorials

24

Some Interesting Articles

Approaches to the Analysis of Survey Data httpwwwilriorgbiometricsTrainingResourcesDocumentsUniversity20of20ReadingGuides

Guides20on20AnalysisApprochAnalysispdf

Likert scales levels of measurement and the ldquolawsrdquo of statistics httpwwwfammedouhsceduresearchFMSRE20Orientation20amp20Handout20Materials

Handouts20520ScienceLikert20Scalespdf

Likert scales how to (ab)use them httpwwwgooglecomurlsa=tamprct=jampq=ampesrc=sampfrm=1ampsource=webampcd=1ampved=0CCoQFjAAampurl=http3A

2F2Fxayimgcom2Fkq2Fgroups2F187517252F1281694392Fname2F1likertscalespdfampei=LU8mUrGbI

MfD2QWe5YHwAQampusg=AFQjCNFQEMmMQbMzNc2go5XL-Cj-wo_rDgampsig2=HhUUy6KVsqqLqqXVxvLkdQampbvm=

bv51495398db2I

Essential elements of questionnaire design and development httpwwwbrighamandwomensorgmedical_professionalscareercfddMentoring20Resourcessurveydesignpdf

25

Thank you

QuestionsComments

RizwanaRehmanvagov

(919) 286-0411 ext 5024

26

For more information program materials and to complete evaluation for CME

credit visit wwwepilepsyvagovStatistics

Page 2: Statistics in Evidence Based Medicine II · Statistics in Evidence Based Medicine II ... Stages of survey research ... commonly used as measures of test-retest

Text Book

Basic and Clinical Biostatistics (2004)

Beth Dawson Robert G Trapp

httpwwwaccessmedicinecomresourceTOCaspxresourceID=62

For more information program materials and to complete evaluation for CME credit visit

wwwepilepsyvagovStatistics

2

Audio Information Dial 1-855-767-1051 Conference ID 61304911

Overview

Stages of survey research

Types of surveys

Questionnaire bias

Measurement data types

Reliability amp validity

Sampling and sample size

Analysis of survey data

3

Stages of the Survey Research Process

4

1 Identifying the focus of the study and method of research

2 Determining the research schedule and budget

3 Establishing an information base

4 Determining the sampling frame

5 Determining the sample size and sample selection procedures

Designing and Conducing Survey Research a Comprehensive Guide Luis M Rea amp Richard A Parker

Stages of the Survey Research Process

6 Designing the survey instrument

7 Pretesting the survey instrument

8 Selecting and training interviewers

9 Implementing the survey

10 Coding the completed questionnaires and computerizing the data

11 Analyzing the data and preparing the final report

5 Designing and Conducing Survey Research a Comprehensive Guide Luis M Rea amp Richard A Parker

Types of Sample Survey Research

Mail out surveys

Web based surveys

The telephone survey

In-person interviews

Intercept surveys

6

Open-ended vs Closed Questions

7

Basic and Clinical Biostatistics Chapter 11

Questionnaire Bias

Effect of varying question wording

Different interpretation of respondents from the questioner

Respondents donrsquot understand the question

Setting in which question is asked

8

Levels of Measurement

9

Nominal

Ordinal

Likert scales

Interval

Likert Scale

Likert scales use fixed choice response formats

Designed to measure attitudes or opinion

A Likert type scale assumes that the strengthintensity of experience is linear

Most often five to seven choices including a neutral choice are offered

Common parametric tests are generally used

but analysis should depend upon the research question

10

Reliability of a Survey

11

Reliability is a statistical measure of how

reproducible the survey instrument data

are

Two survey instruments with same outcome

of interest may have different reliabilities

Types of Reliabilities

Test retest

Alternate form

Internal consistency

Intra-observer

Inter-observer

12

Test-retest Reliability

Test-retest reliability is measured by having the same respondents complete a survey at two different points in time

Test-retest reliability is measured by intra-class correlation coefficient

13

22

2

Bewteenwithin

Bewteen

SS

SICC

Swithin and SBetween are the within and between subject standard deviations

Example of Test-retest Reliability

Juniper et al (1996) evaluated the pediatric Asthma Quality of Life Questionnaire (PAQLQ) by examining reliability in children aged 7 to 17 who had stable asthma

14

Item Within subject SD

Between subject SD

ICC

Overall QoL 017 073 095

Symptoms 022 084 093

Activities 042 096 084

Emotions 023 064 089

Medical Statistics A Text Book for the Health Sciences Michael Campbell David Machin amp Stephen Walters

Internal Consistency Reliability

A set of questions measure different aspects of one concept

Cronbachrsquos coefficient alpha (αChronbach) measures internal consistency reliability among a group of items to form a single scale

15

2

2

11 T

i

Chronbachs

s

k

k

k=number of items si=standard deviation of the ith item sT = standard deviation of the sum score T of all the items

Example of Internal Consistency

McKinley at al (1997) used a questionnaire to measure patient satisfaction with out-of ndashhours calls made to general practitioners They measured aspects such as satisfaction management and the doctorrsquos attitude They found values of αChronbach for each score ranging from 061 to 088 They concluded their questionnaire had satisfactory internal consistency

16

Medical Statistics A Text Book for the Health Sciences Michael Campbell David Machin amp Stephen Walters

Inter-observer(Inter-rater) Reliability

Inter-observer reliability provides a measure of how well two or more evaluators agree in their assessment of a variable A survey can be compared with a gold standard by using Cohenrsquos Kappa (k)

17

Chance

ChanceAgree

P

PP

1

PAgree= Estimated Probability the observerrsquos agree PChance= Proportion expected to show agreement by chance

Example of Inter-observer Reliability

18

Clamp and Kendrick (1998) used a telephone survey asking about safety to 165 families with children under age 5 years who had been involved in randomized trial concerning safety advice They chose a random sample of 20 families from the survey who then received a home visit 2 weeks later to measure the consistency of the response to the questions posed The investigators found that for most questions a high kappa value ( gt059) was obtained

Medical Statistics A Text Book for the Health Sciences Michael Campbell David Machin amp Stephen Walters

Validity of a Survey

A survey is valid when it measures what it is supposed to measure

Content

Face

Criterion

Construct

19

Use of Correlation Coefficient for Reliability and Validity

Pearson correlation coefficients r are commonly used as measures of test-retest reliability inter-observer reliability and concurrent validity

Results could be misleading due to systematic biases

20

Determining the Sample Size

Survey should have enough power to find an effect if there exists one

Q What sample size would be enough to generalize results

Ans Determination of sample size depends upon study outcomes

Researcher must consider

Level of confidence

Margin of error 21

Survey Sampling Methods

Simple random sampling

Systematic random sampling

Stratified sampling

Cluster sampling

Convenience sampling

22

Analyzing Survey Data Descriptive Statistics

Measure of central tendency

Measure of dispersion

Inferential Statistics

Parametric amp non-parametric tests

Chi square

ANOVA

Regression amp correlation

More advanced techniques

23

Openstat A free software

Download

httpwwwstatprograms4ucomOpenStatMainhtm

Tutorial

httpmmc2geofisicaunammxcursosgeoestSoftwareOpenStatATextBookpdf

Youtube tutorials

24

Some Interesting Articles

Approaches to the Analysis of Survey Data httpwwwilriorgbiometricsTrainingResourcesDocumentsUniversity20of20ReadingGuides

Guides20on20AnalysisApprochAnalysispdf

Likert scales levels of measurement and the ldquolawsrdquo of statistics httpwwwfammedouhsceduresearchFMSRE20Orientation20amp20Handout20Materials

Handouts20520ScienceLikert20Scalespdf

Likert scales how to (ab)use them httpwwwgooglecomurlsa=tamprct=jampq=ampesrc=sampfrm=1ampsource=webampcd=1ampved=0CCoQFjAAampurl=http3A

2F2Fxayimgcom2Fkq2Fgroups2F187517252F1281694392Fname2F1likertscalespdfampei=LU8mUrGbI

MfD2QWe5YHwAQampusg=AFQjCNFQEMmMQbMzNc2go5XL-Cj-wo_rDgampsig2=HhUUy6KVsqqLqqXVxvLkdQampbvm=

bv51495398db2I

Essential elements of questionnaire design and development httpwwwbrighamandwomensorgmedical_professionalscareercfddMentoring20Resourcessurveydesignpdf

25

Thank you

QuestionsComments

RizwanaRehmanvagov

(919) 286-0411 ext 5024

26

For more information program materials and to complete evaluation for CME

credit visit wwwepilepsyvagovStatistics

Page 3: Statistics in Evidence Based Medicine II · Statistics in Evidence Based Medicine II ... Stages of survey research ... commonly used as measures of test-retest

Overview

Stages of survey research

Types of surveys

Questionnaire bias

Measurement data types

Reliability amp validity

Sampling and sample size

Analysis of survey data

3

Stages of the Survey Research Process

4

1 Identifying the focus of the study and method of research

2 Determining the research schedule and budget

3 Establishing an information base

4 Determining the sampling frame

5 Determining the sample size and sample selection procedures

Designing and Conducing Survey Research a Comprehensive Guide Luis M Rea amp Richard A Parker

Stages of the Survey Research Process

6 Designing the survey instrument

7 Pretesting the survey instrument

8 Selecting and training interviewers

9 Implementing the survey

10 Coding the completed questionnaires and computerizing the data

11 Analyzing the data and preparing the final report

5 Designing and Conducing Survey Research a Comprehensive Guide Luis M Rea amp Richard A Parker

Types of Sample Survey Research

Mail out surveys

Web based surveys

The telephone survey

In-person interviews

Intercept surveys

6

Open-ended vs Closed Questions

7

Basic and Clinical Biostatistics Chapter 11

Questionnaire Bias

Effect of varying question wording

Different interpretation of respondents from the questioner

Respondents donrsquot understand the question

Setting in which question is asked

8

Levels of Measurement

9

Nominal

Ordinal

Likert scales

Interval

Likert Scale

Likert scales use fixed choice response formats

Designed to measure attitudes or opinion

A Likert type scale assumes that the strengthintensity of experience is linear

Most often five to seven choices including a neutral choice are offered

Common parametric tests are generally used

but analysis should depend upon the research question

10

Reliability of a Survey

11

Reliability is a statistical measure of how

reproducible the survey instrument data

are

Two survey instruments with same outcome

of interest may have different reliabilities

Types of Reliabilities

Test retest

Alternate form

Internal consistency

Intra-observer

Inter-observer

12

Test-retest Reliability

Test-retest reliability is measured by having the same respondents complete a survey at two different points in time

Test-retest reliability is measured by intra-class correlation coefficient

13

22

2

Bewteenwithin

Bewteen

SS

SICC

Swithin and SBetween are the within and between subject standard deviations

Example of Test-retest Reliability

Juniper et al (1996) evaluated the pediatric Asthma Quality of Life Questionnaire (PAQLQ) by examining reliability in children aged 7 to 17 who had stable asthma

14

Item Within subject SD

Between subject SD

ICC

Overall QoL 017 073 095

Symptoms 022 084 093

Activities 042 096 084

Emotions 023 064 089

Medical Statistics A Text Book for the Health Sciences Michael Campbell David Machin amp Stephen Walters

Internal Consistency Reliability

A set of questions measure different aspects of one concept

Cronbachrsquos coefficient alpha (αChronbach) measures internal consistency reliability among a group of items to form a single scale

15

2

2

11 T

i

Chronbachs

s

k

k

k=number of items si=standard deviation of the ith item sT = standard deviation of the sum score T of all the items

Example of Internal Consistency

McKinley at al (1997) used a questionnaire to measure patient satisfaction with out-of ndashhours calls made to general practitioners They measured aspects such as satisfaction management and the doctorrsquos attitude They found values of αChronbach for each score ranging from 061 to 088 They concluded their questionnaire had satisfactory internal consistency

16

Medical Statistics A Text Book for the Health Sciences Michael Campbell David Machin amp Stephen Walters

Inter-observer(Inter-rater) Reliability

Inter-observer reliability provides a measure of how well two or more evaluators agree in their assessment of a variable A survey can be compared with a gold standard by using Cohenrsquos Kappa (k)

17

Chance

ChanceAgree

P

PP

1

PAgree= Estimated Probability the observerrsquos agree PChance= Proportion expected to show agreement by chance

Example of Inter-observer Reliability

18

Clamp and Kendrick (1998) used a telephone survey asking about safety to 165 families with children under age 5 years who had been involved in randomized trial concerning safety advice They chose a random sample of 20 families from the survey who then received a home visit 2 weeks later to measure the consistency of the response to the questions posed The investigators found that for most questions a high kappa value ( gt059) was obtained

Medical Statistics A Text Book for the Health Sciences Michael Campbell David Machin amp Stephen Walters

Validity of a Survey

A survey is valid when it measures what it is supposed to measure

Content

Face

Criterion

Construct

19

Use of Correlation Coefficient for Reliability and Validity

Pearson correlation coefficients r are commonly used as measures of test-retest reliability inter-observer reliability and concurrent validity

Results could be misleading due to systematic biases

20

Determining the Sample Size

Survey should have enough power to find an effect if there exists one

Q What sample size would be enough to generalize results

Ans Determination of sample size depends upon study outcomes

Researcher must consider

Level of confidence

Margin of error 21

Survey Sampling Methods

Simple random sampling

Systematic random sampling

Stratified sampling

Cluster sampling

Convenience sampling

22

Analyzing Survey Data Descriptive Statistics

Measure of central tendency

Measure of dispersion

Inferential Statistics

Parametric amp non-parametric tests

Chi square

ANOVA

Regression amp correlation

More advanced techniques

23

Openstat A free software

Download

httpwwwstatprograms4ucomOpenStatMainhtm

Tutorial

httpmmc2geofisicaunammxcursosgeoestSoftwareOpenStatATextBookpdf

Youtube tutorials

24

Some Interesting Articles

Approaches to the Analysis of Survey Data httpwwwilriorgbiometricsTrainingResourcesDocumentsUniversity20of20ReadingGuides

Guides20on20AnalysisApprochAnalysispdf

Likert scales levels of measurement and the ldquolawsrdquo of statistics httpwwwfammedouhsceduresearchFMSRE20Orientation20amp20Handout20Materials

Handouts20520ScienceLikert20Scalespdf

Likert scales how to (ab)use them httpwwwgooglecomurlsa=tamprct=jampq=ampesrc=sampfrm=1ampsource=webampcd=1ampved=0CCoQFjAAampurl=http3A

2F2Fxayimgcom2Fkq2Fgroups2F187517252F1281694392Fname2F1likertscalespdfampei=LU8mUrGbI

MfD2QWe5YHwAQampusg=AFQjCNFQEMmMQbMzNc2go5XL-Cj-wo_rDgampsig2=HhUUy6KVsqqLqqXVxvLkdQampbvm=

bv51495398db2I

Essential elements of questionnaire design and development httpwwwbrighamandwomensorgmedical_professionalscareercfddMentoring20Resourcessurveydesignpdf

25

Thank you

QuestionsComments

RizwanaRehmanvagov

(919) 286-0411 ext 5024

26

For more information program materials and to complete evaluation for CME

credit visit wwwepilepsyvagovStatistics

Page 4: Statistics in Evidence Based Medicine II · Statistics in Evidence Based Medicine II ... Stages of survey research ... commonly used as measures of test-retest

Stages of the Survey Research Process

4

1 Identifying the focus of the study and method of research

2 Determining the research schedule and budget

3 Establishing an information base

4 Determining the sampling frame

5 Determining the sample size and sample selection procedures

Designing and Conducing Survey Research a Comprehensive Guide Luis M Rea amp Richard A Parker

Stages of the Survey Research Process

6 Designing the survey instrument

7 Pretesting the survey instrument

8 Selecting and training interviewers

9 Implementing the survey

10 Coding the completed questionnaires and computerizing the data

11 Analyzing the data and preparing the final report

5 Designing and Conducing Survey Research a Comprehensive Guide Luis M Rea amp Richard A Parker

Types of Sample Survey Research

Mail out surveys

Web based surveys

The telephone survey

In-person interviews

Intercept surveys

6

Open-ended vs Closed Questions

7

Basic and Clinical Biostatistics Chapter 11

Questionnaire Bias

Effect of varying question wording

Different interpretation of respondents from the questioner

Respondents donrsquot understand the question

Setting in which question is asked

8

Levels of Measurement

9

Nominal

Ordinal

Likert scales

Interval

Likert Scale

Likert scales use fixed choice response formats

Designed to measure attitudes or opinion

A Likert type scale assumes that the strengthintensity of experience is linear

Most often five to seven choices including a neutral choice are offered

Common parametric tests are generally used

but analysis should depend upon the research question

10

Reliability of a Survey

11

Reliability is a statistical measure of how

reproducible the survey instrument data

are

Two survey instruments with same outcome

of interest may have different reliabilities

Types of Reliabilities

Test retest

Alternate form

Internal consistency

Intra-observer

Inter-observer

12

Test-retest Reliability

Test-retest reliability is measured by having the same respondents complete a survey at two different points in time

Test-retest reliability is measured by intra-class correlation coefficient

13

22

2

Bewteenwithin

Bewteen

SS

SICC

Swithin and SBetween are the within and between subject standard deviations

Example of Test-retest Reliability

Juniper et al (1996) evaluated the pediatric Asthma Quality of Life Questionnaire (PAQLQ) by examining reliability in children aged 7 to 17 who had stable asthma

14

Item Within subject SD

Between subject SD

ICC

Overall QoL 017 073 095

Symptoms 022 084 093

Activities 042 096 084

Emotions 023 064 089

Medical Statistics A Text Book for the Health Sciences Michael Campbell David Machin amp Stephen Walters

Internal Consistency Reliability

A set of questions measure different aspects of one concept

Cronbachrsquos coefficient alpha (αChronbach) measures internal consistency reliability among a group of items to form a single scale

15

2

2

11 T

i

Chronbachs

s

k

k

k=number of items si=standard deviation of the ith item sT = standard deviation of the sum score T of all the items

Example of Internal Consistency

McKinley at al (1997) used a questionnaire to measure patient satisfaction with out-of ndashhours calls made to general practitioners They measured aspects such as satisfaction management and the doctorrsquos attitude They found values of αChronbach for each score ranging from 061 to 088 They concluded their questionnaire had satisfactory internal consistency

16

Medical Statistics A Text Book for the Health Sciences Michael Campbell David Machin amp Stephen Walters

Inter-observer(Inter-rater) Reliability

Inter-observer reliability provides a measure of how well two or more evaluators agree in their assessment of a variable A survey can be compared with a gold standard by using Cohenrsquos Kappa (k)

17

Chance

ChanceAgree

P

PP

1

PAgree= Estimated Probability the observerrsquos agree PChance= Proportion expected to show agreement by chance

Example of Inter-observer Reliability

18

Clamp and Kendrick (1998) used a telephone survey asking about safety to 165 families with children under age 5 years who had been involved in randomized trial concerning safety advice They chose a random sample of 20 families from the survey who then received a home visit 2 weeks later to measure the consistency of the response to the questions posed The investigators found that for most questions a high kappa value ( gt059) was obtained

Medical Statistics A Text Book for the Health Sciences Michael Campbell David Machin amp Stephen Walters

Validity of a Survey

A survey is valid when it measures what it is supposed to measure

Content

Face

Criterion

Construct

19

Use of Correlation Coefficient for Reliability and Validity

Pearson correlation coefficients r are commonly used as measures of test-retest reliability inter-observer reliability and concurrent validity

Results could be misleading due to systematic biases

20

Determining the Sample Size

Survey should have enough power to find an effect if there exists one

Q What sample size would be enough to generalize results

Ans Determination of sample size depends upon study outcomes

Researcher must consider

Level of confidence

Margin of error 21

Survey Sampling Methods

Simple random sampling

Systematic random sampling

Stratified sampling

Cluster sampling

Convenience sampling

22

Analyzing Survey Data Descriptive Statistics

Measure of central tendency

Measure of dispersion

Inferential Statistics

Parametric amp non-parametric tests

Chi square

ANOVA

Regression amp correlation

More advanced techniques

23

Openstat A free software

Download

httpwwwstatprograms4ucomOpenStatMainhtm

Tutorial

httpmmc2geofisicaunammxcursosgeoestSoftwareOpenStatATextBookpdf

Youtube tutorials

24

Some Interesting Articles

Approaches to the Analysis of Survey Data httpwwwilriorgbiometricsTrainingResourcesDocumentsUniversity20of20ReadingGuides

Guides20on20AnalysisApprochAnalysispdf

Likert scales levels of measurement and the ldquolawsrdquo of statistics httpwwwfammedouhsceduresearchFMSRE20Orientation20amp20Handout20Materials

Handouts20520ScienceLikert20Scalespdf

Likert scales how to (ab)use them httpwwwgooglecomurlsa=tamprct=jampq=ampesrc=sampfrm=1ampsource=webampcd=1ampved=0CCoQFjAAampurl=http3A

2F2Fxayimgcom2Fkq2Fgroups2F187517252F1281694392Fname2F1likertscalespdfampei=LU8mUrGbI

MfD2QWe5YHwAQampusg=AFQjCNFQEMmMQbMzNc2go5XL-Cj-wo_rDgampsig2=HhUUy6KVsqqLqqXVxvLkdQampbvm=

bv51495398db2I

Essential elements of questionnaire design and development httpwwwbrighamandwomensorgmedical_professionalscareercfddMentoring20Resourcessurveydesignpdf

25

Thank you

QuestionsComments

RizwanaRehmanvagov

(919) 286-0411 ext 5024

26

For more information program materials and to complete evaluation for CME

credit visit wwwepilepsyvagovStatistics

Page 5: Statistics in Evidence Based Medicine II · Statistics in Evidence Based Medicine II ... Stages of survey research ... commonly used as measures of test-retest

Stages of the Survey Research Process

6 Designing the survey instrument

7 Pretesting the survey instrument

8 Selecting and training interviewers

9 Implementing the survey

10 Coding the completed questionnaires and computerizing the data

11 Analyzing the data and preparing the final report

5 Designing and Conducing Survey Research a Comprehensive Guide Luis M Rea amp Richard A Parker

Types of Sample Survey Research

Mail out surveys

Web based surveys

The telephone survey

In-person interviews

Intercept surveys

6

Open-ended vs Closed Questions

7

Basic and Clinical Biostatistics Chapter 11

Questionnaire Bias

Effect of varying question wording

Different interpretation of respondents from the questioner

Respondents donrsquot understand the question

Setting in which question is asked

8

Levels of Measurement

9

Nominal

Ordinal

Likert scales

Interval

Likert Scale

Likert scales use fixed choice response formats

Designed to measure attitudes or opinion

A Likert type scale assumes that the strengthintensity of experience is linear

Most often five to seven choices including a neutral choice are offered

Common parametric tests are generally used

but analysis should depend upon the research question

10

Reliability of a Survey

11

Reliability is a statistical measure of how

reproducible the survey instrument data

are

Two survey instruments with same outcome

of interest may have different reliabilities

Types of Reliabilities

Test retest

Alternate form

Internal consistency

Intra-observer

Inter-observer

12

Test-retest Reliability

Test-retest reliability is measured by having the same respondents complete a survey at two different points in time

Test-retest reliability is measured by intra-class correlation coefficient

13

22

2

Bewteenwithin

Bewteen

SS

SICC

Swithin and SBetween are the within and between subject standard deviations

Example of Test-retest Reliability

Juniper et al (1996) evaluated the pediatric Asthma Quality of Life Questionnaire (PAQLQ) by examining reliability in children aged 7 to 17 who had stable asthma

14

Item Within subject SD

Between subject SD

ICC

Overall QoL 017 073 095

Symptoms 022 084 093

Activities 042 096 084

Emotions 023 064 089

Medical Statistics A Text Book for the Health Sciences Michael Campbell David Machin amp Stephen Walters

Internal Consistency Reliability

A set of questions measure different aspects of one concept

Cronbachrsquos coefficient alpha (αChronbach) measures internal consistency reliability among a group of items to form a single scale

15

2

2

11 T

i

Chronbachs

s

k

k

k=number of items si=standard deviation of the ith item sT = standard deviation of the sum score T of all the items

Example of Internal Consistency

McKinley at al (1997) used a questionnaire to measure patient satisfaction with out-of ndashhours calls made to general practitioners They measured aspects such as satisfaction management and the doctorrsquos attitude They found values of αChronbach for each score ranging from 061 to 088 They concluded their questionnaire had satisfactory internal consistency

16

Medical Statistics A Text Book for the Health Sciences Michael Campbell David Machin amp Stephen Walters

Inter-observer(Inter-rater) Reliability

Inter-observer reliability provides a measure of how well two or more evaluators agree in their assessment of a variable A survey can be compared with a gold standard by using Cohenrsquos Kappa (k)

17

Chance

ChanceAgree

P

PP

1

PAgree= Estimated Probability the observerrsquos agree PChance= Proportion expected to show agreement by chance

Example of Inter-observer Reliability

18

Clamp and Kendrick (1998) used a telephone survey asking about safety to 165 families with children under age 5 years who had been involved in randomized trial concerning safety advice They chose a random sample of 20 families from the survey who then received a home visit 2 weeks later to measure the consistency of the response to the questions posed The investigators found that for most questions a high kappa value ( gt059) was obtained

Medical Statistics A Text Book for the Health Sciences Michael Campbell David Machin amp Stephen Walters

Validity of a Survey

A survey is valid when it measures what it is supposed to measure

Content

Face

Criterion

Construct

19

Use of Correlation Coefficient for Reliability and Validity

Pearson correlation coefficients r are commonly used as measures of test-retest reliability inter-observer reliability and concurrent validity

Results could be misleading due to systematic biases

20

Determining the Sample Size

Survey should have enough power to find an effect if there exists one

Q What sample size would be enough to generalize results

Ans Determination of sample size depends upon study outcomes

Researcher must consider

Level of confidence

Margin of error 21

Survey Sampling Methods

Simple random sampling

Systematic random sampling

Stratified sampling

Cluster sampling

Convenience sampling

22

Analyzing Survey Data Descriptive Statistics

Measure of central tendency

Measure of dispersion

Inferential Statistics

Parametric amp non-parametric tests

Chi square

ANOVA

Regression amp correlation

More advanced techniques

23

Openstat A free software

Download

httpwwwstatprograms4ucomOpenStatMainhtm

Tutorial

httpmmc2geofisicaunammxcursosgeoestSoftwareOpenStatATextBookpdf

Youtube tutorials

24

Some Interesting Articles

Approaches to the Analysis of Survey Data httpwwwilriorgbiometricsTrainingResourcesDocumentsUniversity20of20ReadingGuides

Guides20on20AnalysisApprochAnalysispdf

Likert scales levels of measurement and the ldquolawsrdquo of statistics httpwwwfammedouhsceduresearchFMSRE20Orientation20amp20Handout20Materials

Handouts20520ScienceLikert20Scalespdf

Likert scales how to (ab)use them httpwwwgooglecomurlsa=tamprct=jampq=ampesrc=sampfrm=1ampsource=webampcd=1ampved=0CCoQFjAAampurl=http3A

2F2Fxayimgcom2Fkq2Fgroups2F187517252F1281694392Fname2F1likertscalespdfampei=LU8mUrGbI

MfD2QWe5YHwAQampusg=AFQjCNFQEMmMQbMzNc2go5XL-Cj-wo_rDgampsig2=HhUUy6KVsqqLqqXVxvLkdQampbvm=

bv51495398db2I

Essential elements of questionnaire design and development httpwwwbrighamandwomensorgmedical_professionalscareercfddMentoring20Resourcessurveydesignpdf

25

Thank you

QuestionsComments

RizwanaRehmanvagov

(919) 286-0411 ext 5024

26

For more information program materials and to complete evaluation for CME

credit visit wwwepilepsyvagovStatistics

Page 6: Statistics in Evidence Based Medicine II · Statistics in Evidence Based Medicine II ... Stages of survey research ... commonly used as measures of test-retest

Types of Sample Survey Research

Mail out surveys

Web based surveys

The telephone survey

In-person interviews

Intercept surveys

6

Open-ended vs Closed Questions

7

Basic and Clinical Biostatistics Chapter 11

Questionnaire Bias

Effect of varying question wording

Different interpretation of respondents from the questioner

Respondents donrsquot understand the question

Setting in which question is asked

8

Levels of Measurement

9

Nominal

Ordinal

Likert scales

Interval

Likert Scale

Likert scales use fixed choice response formats

Designed to measure attitudes or opinion

A Likert type scale assumes that the strengthintensity of experience is linear

Most often five to seven choices including a neutral choice are offered

Common parametric tests are generally used

but analysis should depend upon the research question

10

Reliability of a Survey

11

Reliability is a statistical measure of how

reproducible the survey instrument data

are

Two survey instruments with same outcome

of interest may have different reliabilities

Types of Reliabilities

Test retest

Alternate form

Internal consistency

Intra-observer

Inter-observer

12

Test-retest Reliability

Test-retest reliability is measured by having the same respondents complete a survey at two different points in time

Test-retest reliability is measured by intra-class correlation coefficient

13

22

2

Bewteenwithin

Bewteen

SS

SICC

Swithin and SBetween are the within and between subject standard deviations

Example of Test-retest Reliability

Juniper et al (1996) evaluated the pediatric Asthma Quality of Life Questionnaire (PAQLQ) by examining reliability in children aged 7 to 17 who had stable asthma

14

Item Within subject SD

Between subject SD

ICC

Overall QoL 017 073 095

Symptoms 022 084 093

Activities 042 096 084

Emotions 023 064 089

Medical Statistics A Text Book for the Health Sciences Michael Campbell David Machin amp Stephen Walters

Internal Consistency Reliability

A set of questions measure different aspects of one concept

Cronbachrsquos coefficient alpha (αChronbach) measures internal consistency reliability among a group of items to form a single scale

15

2

2

11 T

i

Chronbachs

s

k

k

k=number of items si=standard deviation of the ith item sT = standard deviation of the sum score T of all the items

Example of Internal Consistency

McKinley at al (1997) used a questionnaire to measure patient satisfaction with out-of ndashhours calls made to general practitioners They measured aspects such as satisfaction management and the doctorrsquos attitude They found values of αChronbach for each score ranging from 061 to 088 They concluded their questionnaire had satisfactory internal consistency

16

Medical Statistics A Text Book for the Health Sciences Michael Campbell David Machin amp Stephen Walters

Inter-observer(Inter-rater) Reliability

Inter-observer reliability provides a measure of how well two or more evaluators agree in their assessment of a variable A survey can be compared with a gold standard by using Cohenrsquos Kappa (k)

17

Chance

ChanceAgree

P

PP

1

PAgree= Estimated Probability the observerrsquos agree PChance= Proportion expected to show agreement by chance

Example of Inter-observer Reliability

18

Clamp and Kendrick (1998) used a telephone survey asking about safety to 165 families with children under age 5 years who had been involved in randomized trial concerning safety advice They chose a random sample of 20 families from the survey who then received a home visit 2 weeks later to measure the consistency of the response to the questions posed The investigators found that for most questions a high kappa value ( gt059) was obtained

Medical Statistics A Text Book for the Health Sciences Michael Campbell David Machin amp Stephen Walters

Validity of a Survey

A survey is valid when it measures what it is supposed to measure

Content

Face

Criterion

Construct

19

Use of Correlation Coefficient for Reliability and Validity

Pearson correlation coefficients r are commonly used as measures of test-retest reliability inter-observer reliability and concurrent validity

Results could be misleading due to systematic biases

20

Determining the Sample Size

Survey should have enough power to find an effect if there exists one

Q What sample size would be enough to generalize results

Ans Determination of sample size depends upon study outcomes

Researcher must consider

Level of confidence

Margin of error 21

Survey Sampling Methods

Simple random sampling

Systematic random sampling

Stratified sampling

Cluster sampling

Convenience sampling

22

Analyzing Survey Data Descriptive Statistics

Measure of central tendency

Measure of dispersion

Inferential Statistics

Parametric amp non-parametric tests

Chi square

ANOVA

Regression amp correlation

More advanced techniques

23

Openstat A free software

Download

httpwwwstatprograms4ucomOpenStatMainhtm

Tutorial

httpmmc2geofisicaunammxcursosgeoestSoftwareOpenStatATextBookpdf

Youtube tutorials

24

Some Interesting Articles

Approaches to the Analysis of Survey Data httpwwwilriorgbiometricsTrainingResourcesDocumentsUniversity20of20ReadingGuides

Guides20on20AnalysisApprochAnalysispdf

Likert scales levels of measurement and the ldquolawsrdquo of statistics httpwwwfammedouhsceduresearchFMSRE20Orientation20amp20Handout20Materials

Handouts20520ScienceLikert20Scalespdf

Likert scales how to (ab)use them httpwwwgooglecomurlsa=tamprct=jampq=ampesrc=sampfrm=1ampsource=webampcd=1ampved=0CCoQFjAAampurl=http3A

2F2Fxayimgcom2Fkq2Fgroups2F187517252F1281694392Fname2F1likertscalespdfampei=LU8mUrGbI

MfD2QWe5YHwAQampusg=AFQjCNFQEMmMQbMzNc2go5XL-Cj-wo_rDgampsig2=HhUUy6KVsqqLqqXVxvLkdQampbvm=

bv51495398db2I

Essential elements of questionnaire design and development httpwwwbrighamandwomensorgmedical_professionalscareercfddMentoring20Resourcessurveydesignpdf

25

Thank you

QuestionsComments

RizwanaRehmanvagov

(919) 286-0411 ext 5024

26

For more information program materials and to complete evaluation for CME

credit visit wwwepilepsyvagovStatistics

Page 7: Statistics in Evidence Based Medicine II · Statistics in Evidence Based Medicine II ... Stages of survey research ... commonly used as measures of test-retest

Open-ended vs Closed Questions

7

Basic and Clinical Biostatistics Chapter 11

Questionnaire Bias

Effect of varying question wording

Different interpretation of respondents from the questioner

Respondents donrsquot understand the question

Setting in which question is asked

8

Levels of Measurement

9

Nominal

Ordinal

Likert scales

Interval

Likert Scale

Likert scales use fixed choice response formats

Designed to measure attitudes or opinion

A Likert type scale assumes that the strengthintensity of experience is linear

Most often five to seven choices including a neutral choice are offered

Common parametric tests are generally used

but analysis should depend upon the research question

10

Reliability of a Survey

11

Reliability is a statistical measure of how

reproducible the survey instrument data

are

Two survey instruments with same outcome

of interest may have different reliabilities

Types of Reliabilities

Test retest

Alternate form

Internal consistency

Intra-observer

Inter-observer

12

Test-retest Reliability

Test-retest reliability is measured by having the same respondents complete a survey at two different points in time

Test-retest reliability is measured by intra-class correlation coefficient

13

22

2

Bewteenwithin

Bewteen

SS

SICC

Swithin and SBetween are the within and between subject standard deviations

Example of Test-retest Reliability

Juniper et al (1996) evaluated the pediatric Asthma Quality of Life Questionnaire (PAQLQ) by examining reliability in children aged 7 to 17 who had stable asthma

14

Item Within subject SD

Between subject SD

ICC

Overall QoL 017 073 095

Symptoms 022 084 093

Activities 042 096 084

Emotions 023 064 089

Medical Statistics A Text Book for the Health Sciences Michael Campbell David Machin amp Stephen Walters

Internal Consistency Reliability

A set of questions measure different aspects of one concept

Cronbachrsquos coefficient alpha (αChronbach) measures internal consistency reliability among a group of items to form a single scale

15

2

2

11 T

i

Chronbachs

s

k

k

k=number of items si=standard deviation of the ith item sT = standard deviation of the sum score T of all the items

Example of Internal Consistency

McKinley at al (1997) used a questionnaire to measure patient satisfaction with out-of ndashhours calls made to general practitioners They measured aspects such as satisfaction management and the doctorrsquos attitude They found values of αChronbach for each score ranging from 061 to 088 They concluded their questionnaire had satisfactory internal consistency

16

Medical Statistics A Text Book for the Health Sciences Michael Campbell David Machin amp Stephen Walters

Inter-observer(Inter-rater) Reliability

Inter-observer reliability provides a measure of how well two or more evaluators agree in their assessment of a variable A survey can be compared with a gold standard by using Cohenrsquos Kappa (k)

17

Chance

ChanceAgree

P

PP

1

PAgree= Estimated Probability the observerrsquos agree PChance= Proportion expected to show agreement by chance

Example of Inter-observer Reliability

18

Clamp and Kendrick (1998) used a telephone survey asking about safety to 165 families with children under age 5 years who had been involved in randomized trial concerning safety advice They chose a random sample of 20 families from the survey who then received a home visit 2 weeks later to measure the consistency of the response to the questions posed The investigators found that for most questions a high kappa value ( gt059) was obtained

Medical Statistics A Text Book for the Health Sciences Michael Campbell David Machin amp Stephen Walters

Validity of a Survey

A survey is valid when it measures what it is supposed to measure

Content

Face

Criterion

Construct

19

Use of Correlation Coefficient for Reliability and Validity

Pearson correlation coefficients r are commonly used as measures of test-retest reliability inter-observer reliability and concurrent validity

Results could be misleading due to systematic biases

20

Determining the Sample Size

Survey should have enough power to find an effect if there exists one

Q What sample size would be enough to generalize results

Ans Determination of sample size depends upon study outcomes

Researcher must consider

Level of confidence

Margin of error 21

Survey Sampling Methods

Simple random sampling

Systematic random sampling

Stratified sampling

Cluster sampling

Convenience sampling

22

Analyzing Survey Data Descriptive Statistics

Measure of central tendency

Measure of dispersion

Inferential Statistics

Parametric amp non-parametric tests

Chi square

ANOVA

Regression amp correlation

More advanced techniques

23

Openstat A free software

Download

httpwwwstatprograms4ucomOpenStatMainhtm

Tutorial

httpmmc2geofisicaunammxcursosgeoestSoftwareOpenStatATextBookpdf

Youtube tutorials

24

Some Interesting Articles

Approaches to the Analysis of Survey Data httpwwwilriorgbiometricsTrainingResourcesDocumentsUniversity20of20ReadingGuides

Guides20on20AnalysisApprochAnalysispdf

Likert scales levels of measurement and the ldquolawsrdquo of statistics httpwwwfammedouhsceduresearchFMSRE20Orientation20amp20Handout20Materials

Handouts20520ScienceLikert20Scalespdf

Likert scales how to (ab)use them httpwwwgooglecomurlsa=tamprct=jampq=ampesrc=sampfrm=1ampsource=webampcd=1ampved=0CCoQFjAAampurl=http3A

2F2Fxayimgcom2Fkq2Fgroups2F187517252F1281694392Fname2F1likertscalespdfampei=LU8mUrGbI

MfD2QWe5YHwAQampusg=AFQjCNFQEMmMQbMzNc2go5XL-Cj-wo_rDgampsig2=HhUUy6KVsqqLqqXVxvLkdQampbvm=

bv51495398db2I

Essential elements of questionnaire design and development httpwwwbrighamandwomensorgmedical_professionalscareercfddMentoring20Resourcessurveydesignpdf

25

Thank you

QuestionsComments

RizwanaRehmanvagov

(919) 286-0411 ext 5024

26

For more information program materials and to complete evaluation for CME

credit visit wwwepilepsyvagovStatistics

Page 8: Statistics in Evidence Based Medicine II · Statistics in Evidence Based Medicine II ... Stages of survey research ... commonly used as measures of test-retest

Questionnaire Bias

Effect of varying question wording

Different interpretation of respondents from the questioner

Respondents donrsquot understand the question

Setting in which question is asked

8

Levels of Measurement

9

Nominal

Ordinal

Likert scales

Interval

Likert Scale

Likert scales use fixed choice response formats

Designed to measure attitudes or opinion

A Likert type scale assumes that the strengthintensity of experience is linear

Most often five to seven choices including a neutral choice are offered

Common parametric tests are generally used

but analysis should depend upon the research question

10

Reliability of a Survey

11

Reliability is a statistical measure of how

reproducible the survey instrument data

are

Two survey instruments with same outcome

of interest may have different reliabilities

Types of Reliabilities

Test retest

Alternate form

Internal consistency

Intra-observer

Inter-observer

12

Test-retest Reliability

Test-retest reliability is measured by having the same respondents complete a survey at two different points in time

Test-retest reliability is measured by intra-class correlation coefficient

13

22

2

Bewteenwithin

Bewteen

SS

SICC

Swithin and SBetween are the within and between subject standard deviations

Example of Test-retest Reliability

Juniper et al (1996) evaluated the pediatric Asthma Quality of Life Questionnaire (PAQLQ) by examining reliability in children aged 7 to 17 who had stable asthma

14

Item Within subject SD

Between subject SD

ICC

Overall QoL 017 073 095

Symptoms 022 084 093

Activities 042 096 084

Emotions 023 064 089

Medical Statistics A Text Book for the Health Sciences Michael Campbell David Machin amp Stephen Walters

Internal Consistency Reliability

A set of questions measure different aspects of one concept

Cronbachrsquos coefficient alpha (αChronbach) measures internal consistency reliability among a group of items to form a single scale

15

2

2

11 T

i

Chronbachs

s

k

k

k=number of items si=standard deviation of the ith item sT = standard deviation of the sum score T of all the items

Example of Internal Consistency

McKinley at al (1997) used a questionnaire to measure patient satisfaction with out-of ndashhours calls made to general practitioners They measured aspects such as satisfaction management and the doctorrsquos attitude They found values of αChronbach for each score ranging from 061 to 088 They concluded their questionnaire had satisfactory internal consistency

16

Medical Statistics A Text Book for the Health Sciences Michael Campbell David Machin amp Stephen Walters

Inter-observer(Inter-rater) Reliability

Inter-observer reliability provides a measure of how well two or more evaluators agree in their assessment of a variable A survey can be compared with a gold standard by using Cohenrsquos Kappa (k)

17

Chance

ChanceAgree

P

PP

1

PAgree= Estimated Probability the observerrsquos agree PChance= Proportion expected to show agreement by chance

Example of Inter-observer Reliability

18

Clamp and Kendrick (1998) used a telephone survey asking about safety to 165 families with children under age 5 years who had been involved in randomized trial concerning safety advice They chose a random sample of 20 families from the survey who then received a home visit 2 weeks later to measure the consistency of the response to the questions posed The investigators found that for most questions a high kappa value ( gt059) was obtained

Medical Statistics A Text Book for the Health Sciences Michael Campbell David Machin amp Stephen Walters

Validity of a Survey

A survey is valid when it measures what it is supposed to measure

Content

Face

Criterion

Construct

19

Use of Correlation Coefficient for Reliability and Validity

Pearson correlation coefficients r are commonly used as measures of test-retest reliability inter-observer reliability and concurrent validity

Results could be misleading due to systematic biases

20

Determining the Sample Size

Survey should have enough power to find an effect if there exists one

Q What sample size would be enough to generalize results

Ans Determination of sample size depends upon study outcomes

Researcher must consider

Level of confidence

Margin of error 21

Survey Sampling Methods

Simple random sampling

Systematic random sampling

Stratified sampling

Cluster sampling

Convenience sampling

22

Analyzing Survey Data Descriptive Statistics

Measure of central tendency

Measure of dispersion

Inferential Statistics

Parametric amp non-parametric tests

Chi square

ANOVA

Regression amp correlation

More advanced techniques

23

Openstat A free software

Download

httpwwwstatprograms4ucomOpenStatMainhtm

Tutorial

httpmmc2geofisicaunammxcursosgeoestSoftwareOpenStatATextBookpdf

Youtube tutorials

24

Some Interesting Articles

Approaches to the Analysis of Survey Data httpwwwilriorgbiometricsTrainingResourcesDocumentsUniversity20of20ReadingGuides

Guides20on20AnalysisApprochAnalysispdf

Likert scales levels of measurement and the ldquolawsrdquo of statistics httpwwwfammedouhsceduresearchFMSRE20Orientation20amp20Handout20Materials

Handouts20520ScienceLikert20Scalespdf

Likert scales how to (ab)use them httpwwwgooglecomurlsa=tamprct=jampq=ampesrc=sampfrm=1ampsource=webampcd=1ampved=0CCoQFjAAampurl=http3A

2F2Fxayimgcom2Fkq2Fgroups2F187517252F1281694392Fname2F1likertscalespdfampei=LU8mUrGbI

MfD2QWe5YHwAQampusg=AFQjCNFQEMmMQbMzNc2go5XL-Cj-wo_rDgampsig2=HhUUy6KVsqqLqqXVxvLkdQampbvm=

bv51495398db2I

Essential elements of questionnaire design and development httpwwwbrighamandwomensorgmedical_professionalscareercfddMentoring20Resourcessurveydesignpdf

25

Thank you

QuestionsComments

RizwanaRehmanvagov

(919) 286-0411 ext 5024

26

For more information program materials and to complete evaluation for CME

credit visit wwwepilepsyvagovStatistics

Page 9: Statistics in Evidence Based Medicine II · Statistics in Evidence Based Medicine II ... Stages of survey research ... commonly used as measures of test-retest

Levels of Measurement

9

Nominal

Ordinal

Likert scales

Interval

Likert Scale

Likert scales use fixed choice response formats

Designed to measure attitudes or opinion

A Likert type scale assumes that the strengthintensity of experience is linear

Most often five to seven choices including a neutral choice are offered

Common parametric tests are generally used

but analysis should depend upon the research question

10

Reliability of a Survey

11

Reliability is a statistical measure of how

reproducible the survey instrument data

are

Two survey instruments with same outcome

of interest may have different reliabilities

Types of Reliabilities

Test retest

Alternate form

Internal consistency

Intra-observer

Inter-observer

12

Test-retest Reliability

Test-retest reliability is measured by having the same respondents complete a survey at two different points in time

Test-retest reliability is measured by intra-class correlation coefficient

13

22

2

Bewteenwithin

Bewteen

SS

SICC

Swithin and SBetween are the within and between subject standard deviations

Example of Test-retest Reliability

Juniper et al (1996) evaluated the pediatric Asthma Quality of Life Questionnaire (PAQLQ) by examining reliability in children aged 7 to 17 who had stable asthma

14

Item Within subject SD

Between subject SD

ICC

Overall QoL 017 073 095

Symptoms 022 084 093

Activities 042 096 084

Emotions 023 064 089

Medical Statistics A Text Book for the Health Sciences Michael Campbell David Machin amp Stephen Walters

Internal Consistency Reliability

A set of questions measure different aspects of one concept

Cronbachrsquos coefficient alpha (αChronbach) measures internal consistency reliability among a group of items to form a single scale

15

2

2

11 T

i

Chronbachs

s

k

k

k=number of items si=standard deviation of the ith item sT = standard deviation of the sum score T of all the items

Example of Internal Consistency

McKinley at al (1997) used a questionnaire to measure patient satisfaction with out-of ndashhours calls made to general practitioners They measured aspects such as satisfaction management and the doctorrsquos attitude They found values of αChronbach for each score ranging from 061 to 088 They concluded their questionnaire had satisfactory internal consistency

16

Medical Statistics A Text Book for the Health Sciences Michael Campbell David Machin amp Stephen Walters

Inter-observer(Inter-rater) Reliability

Inter-observer reliability provides a measure of how well two or more evaluators agree in their assessment of a variable A survey can be compared with a gold standard by using Cohenrsquos Kappa (k)

17

Chance

ChanceAgree

P

PP

1

PAgree= Estimated Probability the observerrsquos agree PChance= Proportion expected to show agreement by chance

Example of Inter-observer Reliability

18

Clamp and Kendrick (1998) used a telephone survey asking about safety to 165 families with children under age 5 years who had been involved in randomized trial concerning safety advice They chose a random sample of 20 families from the survey who then received a home visit 2 weeks later to measure the consistency of the response to the questions posed The investigators found that for most questions a high kappa value ( gt059) was obtained

Medical Statistics A Text Book for the Health Sciences Michael Campbell David Machin amp Stephen Walters

Validity of a Survey

A survey is valid when it measures what it is supposed to measure

Content

Face

Criterion

Construct

19

Use of Correlation Coefficient for Reliability and Validity

Pearson correlation coefficients r are commonly used as measures of test-retest reliability inter-observer reliability and concurrent validity

Results could be misleading due to systematic biases

20

Determining the Sample Size

Survey should have enough power to find an effect if there exists one

Q What sample size would be enough to generalize results

Ans Determination of sample size depends upon study outcomes

Researcher must consider

Level of confidence

Margin of error 21

Survey Sampling Methods

Simple random sampling

Systematic random sampling

Stratified sampling

Cluster sampling

Convenience sampling

22

Analyzing Survey Data Descriptive Statistics

Measure of central tendency

Measure of dispersion

Inferential Statistics

Parametric amp non-parametric tests

Chi square

ANOVA

Regression amp correlation

More advanced techniques

23

Openstat A free software

Download

httpwwwstatprograms4ucomOpenStatMainhtm

Tutorial

httpmmc2geofisicaunammxcursosgeoestSoftwareOpenStatATextBookpdf

Youtube tutorials

24

Some Interesting Articles

Approaches to the Analysis of Survey Data httpwwwilriorgbiometricsTrainingResourcesDocumentsUniversity20of20ReadingGuides

Guides20on20AnalysisApprochAnalysispdf

Likert scales levels of measurement and the ldquolawsrdquo of statistics httpwwwfammedouhsceduresearchFMSRE20Orientation20amp20Handout20Materials

Handouts20520ScienceLikert20Scalespdf

Likert scales how to (ab)use them httpwwwgooglecomurlsa=tamprct=jampq=ampesrc=sampfrm=1ampsource=webampcd=1ampved=0CCoQFjAAampurl=http3A

2F2Fxayimgcom2Fkq2Fgroups2F187517252F1281694392Fname2F1likertscalespdfampei=LU8mUrGbI

MfD2QWe5YHwAQampusg=AFQjCNFQEMmMQbMzNc2go5XL-Cj-wo_rDgampsig2=HhUUy6KVsqqLqqXVxvLkdQampbvm=

bv51495398db2I

Essential elements of questionnaire design and development httpwwwbrighamandwomensorgmedical_professionalscareercfddMentoring20Resourcessurveydesignpdf

25

Thank you

QuestionsComments

RizwanaRehmanvagov

(919) 286-0411 ext 5024

26

For more information program materials and to complete evaluation for CME

credit visit wwwepilepsyvagovStatistics

Page 10: Statistics in Evidence Based Medicine II · Statistics in Evidence Based Medicine II ... Stages of survey research ... commonly used as measures of test-retest

Likert Scale

Likert scales use fixed choice response formats

Designed to measure attitudes or opinion

A Likert type scale assumes that the strengthintensity of experience is linear

Most often five to seven choices including a neutral choice are offered

Common parametric tests are generally used

but analysis should depend upon the research question

10

Reliability of a Survey

11

Reliability is a statistical measure of how

reproducible the survey instrument data

are

Two survey instruments with same outcome

of interest may have different reliabilities

Types of Reliabilities

Test retest

Alternate form

Internal consistency

Intra-observer

Inter-observer

12

Test-retest Reliability

Test-retest reliability is measured by having the same respondents complete a survey at two different points in time

Test-retest reliability is measured by intra-class correlation coefficient

13

22

2

Bewteenwithin

Bewteen

SS

SICC

Swithin and SBetween are the within and between subject standard deviations

Example of Test-retest Reliability

Juniper et al (1996) evaluated the pediatric Asthma Quality of Life Questionnaire (PAQLQ) by examining reliability in children aged 7 to 17 who had stable asthma

14

Item Within subject SD

Between subject SD

ICC

Overall QoL 017 073 095

Symptoms 022 084 093

Activities 042 096 084

Emotions 023 064 089

Medical Statistics A Text Book for the Health Sciences Michael Campbell David Machin amp Stephen Walters

Internal Consistency Reliability

A set of questions measure different aspects of one concept

Cronbachrsquos coefficient alpha (αChronbach) measures internal consistency reliability among a group of items to form a single scale

15

2

2

11 T

i

Chronbachs

s

k

k

k=number of items si=standard deviation of the ith item sT = standard deviation of the sum score T of all the items

Example of Internal Consistency

McKinley at al (1997) used a questionnaire to measure patient satisfaction with out-of ndashhours calls made to general practitioners They measured aspects such as satisfaction management and the doctorrsquos attitude They found values of αChronbach for each score ranging from 061 to 088 They concluded their questionnaire had satisfactory internal consistency

16

Medical Statistics A Text Book for the Health Sciences Michael Campbell David Machin amp Stephen Walters

Inter-observer(Inter-rater) Reliability

Inter-observer reliability provides a measure of how well two or more evaluators agree in their assessment of a variable A survey can be compared with a gold standard by using Cohenrsquos Kappa (k)

17

Chance

ChanceAgree

P

PP

1

PAgree= Estimated Probability the observerrsquos agree PChance= Proportion expected to show agreement by chance

Example of Inter-observer Reliability

18

Clamp and Kendrick (1998) used a telephone survey asking about safety to 165 families with children under age 5 years who had been involved in randomized trial concerning safety advice They chose a random sample of 20 families from the survey who then received a home visit 2 weeks later to measure the consistency of the response to the questions posed The investigators found that for most questions a high kappa value ( gt059) was obtained

Medical Statistics A Text Book for the Health Sciences Michael Campbell David Machin amp Stephen Walters

Validity of a Survey

A survey is valid when it measures what it is supposed to measure

Content

Face

Criterion

Construct

19

Use of Correlation Coefficient for Reliability and Validity

Pearson correlation coefficients r are commonly used as measures of test-retest reliability inter-observer reliability and concurrent validity

Results could be misleading due to systematic biases

20

Determining the Sample Size

Survey should have enough power to find an effect if there exists one

Q What sample size would be enough to generalize results

Ans Determination of sample size depends upon study outcomes

Researcher must consider

Level of confidence

Margin of error 21

Survey Sampling Methods

Simple random sampling

Systematic random sampling

Stratified sampling

Cluster sampling

Convenience sampling

22

Analyzing Survey Data Descriptive Statistics

Measure of central tendency

Measure of dispersion

Inferential Statistics

Parametric amp non-parametric tests

Chi square

ANOVA

Regression amp correlation

More advanced techniques

23

Openstat A free software

Download

httpwwwstatprograms4ucomOpenStatMainhtm

Tutorial

httpmmc2geofisicaunammxcursosgeoestSoftwareOpenStatATextBookpdf

Youtube tutorials

24

Some Interesting Articles

Approaches to the Analysis of Survey Data httpwwwilriorgbiometricsTrainingResourcesDocumentsUniversity20of20ReadingGuides

Guides20on20AnalysisApprochAnalysispdf

Likert scales levels of measurement and the ldquolawsrdquo of statistics httpwwwfammedouhsceduresearchFMSRE20Orientation20amp20Handout20Materials

Handouts20520ScienceLikert20Scalespdf

Likert scales how to (ab)use them httpwwwgooglecomurlsa=tamprct=jampq=ampesrc=sampfrm=1ampsource=webampcd=1ampved=0CCoQFjAAampurl=http3A

2F2Fxayimgcom2Fkq2Fgroups2F187517252F1281694392Fname2F1likertscalespdfampei=LU8mUrGbI

MfD2QWe5YHwAQampusg=AFQjCNFQEMmMQbMzNc2go5XL-Cj-wo_rDgampsig2=HhUUy6KVsqqLqqXVxvLkdQampbvm=

bv51495398db2I

Essential elements of questionnaire design and development httpwwwbrighamandwomensorgmedical_professionalscareercfddMentoring20Resourcessurveydesignpdf

25

Thank you

QuestionsComments

RizwanaRehmanvagov

(919) 286-0411 ext 5024

26

For more information program materials and to complete evaluation for CME

credit visit wwwepilepsyvagovStatistics

Page 11: Statistics in Evidence Based Medicine II · Statistics in Evidence Based Medicine II ... Stages of survey research ... commonly used as measures of test-retest

Reliability of a Survey

11

Reliability is a statistical measure of how

reproducible the survey instrument data

are

Two survey instruments with same outcome

of interest may have different reliabilities

Types of Reliabilities

Test retest

Alternate form

Internal consistency

Intra-observer

Inter-observer

12

Test-retest Reliability

Test-retest reliability is measured by having the same respondents complete a survey at two different points in time

Test-retest reliability is measured by intra-class correlation coefficient

13

22

2

Bewteenwithin

Bewteen

SS

SICC

Swithin and SBetween are the within and between subject standard deviations

Example of Test-retest Reliability

Juniper et al (1996) evaluated the pediatric Asthma Quality of Life Questionnaire (PAQLQ) by examining reliability in children aged 7 to 17 who had stable asthma

14

Item Within subject SD

Between subject SD

ICC

Overall QoL 017 073 095

Symptoms 022 084 093

Activities 042 096 084

Emotions 023 064 089

Medical Statistics A Text Book for the Health Sciences Michael Campbell David Machin amp Stephen Walters

Internal Consistency Reliability

A set of questions measure different aspects of one concept

Cronbachrsquos coefficient alpha (αChronbach) measures internal consistency reliability among a group of items to form a single scale

15

2

2

11 T

i

Chronbachs

s

k

k

k=number of items si=standard deviation of the ith item sT = standard deviation of the sum score T of all the items

Example of Internal Consistency

McKinley at al (1997) used a questionnaire to measure patient satisfaction with out-of ndashhours calls made to general practitioners They measured aspects such as satisfaction management and the doctorrsquos attitude They found values of αChronbach for each score ranging from 061 to 088 They concluded their questionnaire had satisfactory internal consistency

16

Medical Statistics A Text Book for the Health Sciences Michael Campbell David Machin amp Stephen Walters

Inter-observer(Inter-rater) Reliability

Inter-observer reliability provides a measure of how well two or more evaluators agree in their assessment of a variable A survey can be compared with a gold standard by using Cohenrsquos Kappa (k)

17

Chance

ChanceAgree

P

PP

1

PAgree= Estimated Probability the observerrsquos agree PChance= Proportion expected to show agreement by chance

Example of Inter-observer Reliability

18

Clamp and Kendrick (1998) used a telephone survey asking about safety to 165 families with children under age 5 years who had been involved in randomized trial concerning safety advice They chose a random sample of 20 families from the survey who then received a home visit 2 weeks later to measure the consistency of the response to the questions posed The investigators found that for most questions a high kappa value ( gt059) was obtained

Medical Statistics A Text Book for the Health Sciences Michael Campbell David Machin amp Stephen Walters

Validity of a Survey

A survey is valid when it measures what it is supposed to measure

Content

Face

Criterion

Construct

19

Use of Correlation Coefficient for Reliability and Validity

Pearson correlation coefficients r are commonly used as measures of test-retest reliability inter-observer reliability and concurrent validity

Results could be misleading due to systematic biases

20

Determining the Sample Size

Survey should have enough power to find an effect if there exists one

Q What sample size would be enough to generalize results

Ans Determination of sample size depends upon study outcomes

Researcher must consider

Level of confidence

Margin of error 21

Survey Sampling Methods

Simple random sampling

Systematic random sampling

Stratified sampling

Cluster sampling

Convenience sampling

22

Analyzing Survey Data Descriptive Statistics

Measure of central tendency

Measure of dispersion

Inferential Statistics

Parametric amp non-parametric tests

Chi square

ANOVA

Regression amp correlation

More advanced techniques

23

Openstat A free software

Download

httpwwwstatprograms4ucomOpenStatMainhtm

Tutorial

httpmmc2geofisicaunammxcursosgeoestSoftwareOpenStatATextBookpdf

Youtube tutorials

24

Some Interesting Articles

Approaches to the Analysis of Survey Data httpwwwilriorgbiometricsTrainingResourcesDocumentsUniversity20of20ReadingGuides

Guides20on20AnalysisApprochAnalysispdf

Likert scales levels of measurement and the ldquolawsrdquo of statistics httpwwwfammedouhsceduresearchFMSRE20Orientation20amp20Handout20Materials

Handouts20520ScienceLikert20Scalespdf

Likert scales how to (ab)use them httpwwwgooglecomurlsa=tamprct=jampq=ampesrc=sampfrm=1ampsource=webampcd=1ampved=0CCoQFjAAampurl=http3A

2F2Fxayimgcom2Fkq2Fgroups2F187517252F1281694392Fname2F1likertscalespdfampei=LU8mUrGbI

MfD2QWe5YHwAQampusg=AFQjCNFQEMmMQbMzNc2go5XL-Cj-wo_rDgampsig2=HhUUy6KVsqqLqqXVxvLkdQampbvm=

bv51495398db2I

Essential elements of questionnaire design and development httpwwwbrighamandwomensorgmedical_professionalscareercfddMentoring20Resourcessurveydesignpdf

25

Thank you

QuestionsComments

RizwanaRehmanvagov

(919) 286-0411 ext 5024

26

For more information program materials and to complete evaluation for CME

credit visit wwwepilepsyvagovStatistics

Page 12: Statistics in Evidence Based Medicine II · Statistics in Evidence Based Medicine II ... Stages of survey research ... commonly used as measures of test-retest

Types of Reliabilities

Test retest

Alternate form

Internal consistency

Intra-observer

Inter-observer

12

Test-retest Reliability

Test-retest reliability is measured by having the same respondents complete a survey at two different points in time

Test-retest reliability is measured by intra-class correlation coefficient

13

22

2

Bewteenwithin

Bewteen

SS

SICC

Swithin and SBetween are the within and between subject standard deviations

Example of Test-retest Reliability

Juniper et al (1996) evaluated the pediatric Asthma Quality of Life Questionnaire (PAQLQ) by examining reliability in children aged 7 to 17 who had stable asthma

14

Item Within subject SD

Between subject SD

ICC

Overall QoL 017 073 095

Symptoms 022 084 093

Activities 042 096 084

Emotions 023 064 089

Medical Statistics A Text Book for the Health Sciences Michael Campbell David Machin amp Stephen Walters

Internal Consistency Reliability

A set of questions measure different aspects of one concept

Cronbachrsquos coefficient alpha (αChronbach) measures internal consistency reliability among a group of items to form a single scale

15

2

2

11 T

i

Chronbachs

s

k

k

k=number of items si=standard deviation of the ith item sT = standard deviation of the sum score T of all the items

Example of Internal Consistency

McKinley at al (1997) used a questionnaire to measure patient satisfaction with out-of ndashhours calls made to general practitioners They measured aspects such as satisfaction management and the doctorrsquos attitude They found values of αChronbach for each score ranging from 061 to 088 They concluded their questionnaire had satisfactory internal consistency

16

Medical Statistics A Text Book for the Health Sciences Michael Campbell David Machin amp Stephen Walters

Inter-observer(Inter-rater) Reliability

Inter-observer reliability provides a measure of how well two or more evaluators agree in their assessment of a variable A survey can be compared with a gold standard by using Cohenrsquos Kappa (k)

17

Chance

ChanceAgree

P

PP

1

PAgree= Estimated Probability the observerrsquos agree PChance= Proportion expected to show agreement by chance

Example of Inter-observer Reliability

18

Clamp and Kendrick (1998) used a telephone survey asking about safety to 165 families with children under age 5 years who had been involved in randomized trial concerning safety advice They chose a random sample of 20 families from the survey who then received a home visit 2 weeks later to measure the consistency of the response to the questions posed The investigators found that for most questions a high kappa value ( gt059) was obtained

Medical Statistics A Text Book for the Health Sciences Michael Campbell David Machin amp Stephen Walters

Validity of a Survey

A survey is valid when it measures what it is supposed to measure

Content

Face

Criterion

Construct

19

Use of Correlation Coefficient for Reliability and Validity

Pearson correlation coefficients r are commonly used as measures of test-retest reliability inter-observer reliability and concurrent validity

Results could be misleading due to systematic biases

20

Determining the Sample Size

Survey should have enough power to find an effect if there exists one

Q What sample size would be enough to generalize results

Ans Determination of sample size depends upon study outcomes

Researcher must consider

Level of confidence

Margin of error 21

Survey Sampling Methods

Simple random sampling

Systematic random sampling

Stratified sampling

Cluster sampling

Convenience sampling

22

Analyzing Survey Data Descriptive Statistics

Measure of central tendency

Measure of dispersion

Inferential Statistics

Parametric amp non-parametric tests

Chi square

ANOVA

Regression amp correlation

More advanced techniques

23

Openstat A free software

Download

httpwwwstatprograms4ucomOpenStatMainhtm

Tutorial

httpmmc2geofisicaunammxcursosgeoestSoftwareOpenStatATextBookpdf

Youtube tutorials

24

Some Interesting Articles

Approaches to the Analysis of Survey Data httpwwwilriorgbiometricsTrainingResourcesDocumentsUniversity20of20ReadingGuides

Guides20on20AnalysisApprochAnalysispdf

Likert scales levels of measurement and the ldquolawsrdquo of statistics httpwwwfammedouhsceduresearchFMSRE20Orientation20amp20Handout20Materials

Handouts20520ScienceLikert20Scalespdf

Likert scales how to (ab)use them httpwwwgooglecomurlsa=tamprct=jampq=ampesrc=sampfrm=1ampsource=webampcd=1ampved=0CCoQFjAAampurl=http3A

2F2Fxayimgcom2Fkq2Fgroups2F187517252F1281694392Fname2F1likertscalespdfampei=LU8mUrGbI

MfD2QWe5YHwAQampusg=AFQjCNFQEMmMQbMzNc2go5XL-Cj-wo_rDgampsig2=HhUUy6KVsqqLqqXVxvLkdQampbvm=

bv51495398db2I

Essential elements of questionnaire design and development httpwwwbrighamandwomensorgmedical_professionalscareercfddMentoring20Resourcessurveydesignpdf

25

Thank you

QuestionsComments

RizwanaRehmanvagov

(919) 286-0411 ext 5024

26

For more information program materials and to complete evaluation for CME

credit visit wwwepilepsyvagovStatistics

Page 13: Statistics in Evidence Based Medicine II · Statistics in Evidence Based Medicine II ... Stages of survey research ... commonly used as measures of test-retest

Test-retest Reliability

Test-retest reliability is measured by having the same respondents complete a survey at two different points in time

Test-retest reliability is measured by intra-class correlation coefficient

13

22

2

Bewteenwithin

Bewteen

SS

SICC

Swithin and SBetween are the within and between subject standard deviations

Example of Test-retest Reliability

Juniper et al (1996) evaluated the pediatric Asthma Quality of Life Questionnaire (PAQLQ) by examining reliability in children aged 7 to 17 who had stable asthma

14

Item Within subject SD

Between subject SD

ICC

Overall QoL 017 073 095

Symptoms 022 084 093

Activities 042 096 084

Emotions 023 064 089

Medical Statistics A Text Book for the Health Sciences Michael Campbell David Machin amp Stephen Walters

Internal Consistency Reliability

A set of questions measure different aspects of one concept

Cronbachrsquos coefficient alpha (αChronbach) measures internal consistency reliability among a group of items to form a single scale

15

2

2

11 T

i

Chronbachs

s

k

k

k=number of items si=standard deviation of the ith item sT = standard deviation of the sum score T of all the items

Example of Internal Consistency

McKinley at al (1997) used a questionnaire to measure patient satisfaction with out-of ndashhours calls made to general practitioners They measured aspects such as satisfaction management and the doctorrsquos attitude They found values of αChronbach for each score ranging from 061 to 088 They concluded their questionnaire had satisfactory internal consistency

16

Medical Statistics A Text Book for the Health Sciences Michael Campbell David Machin amp Stephen Walters

Inter-observer(Inter-rater) Reliability

Inter-observer reliability provides a measure of how well two or more evaluators agree in their assessment of a variable A survey can be compared with a gold standard by using Cohenrsquos Kappa (k)

17

Chance

ChanceAgree

P

PP

1

PAgree= Estimated Probability the observerrsquos agree PChance= Proportion expected to show agreement by chance

Example of Inter-observer Reliability

18

Clamp and Kendrick (1998) used a telephone survey asking about safety to 165 families with children under age 5 years who had been involved in randomized trial concerning safety advice They chose a random sample of 20 families from the survey who then received a home visit 2 weeks later to measure the consistency of the response to the questions posed The investigators found that for most questions a high kappa value ( gt059) was obtained

Medical Statistics A Text Book for the Health Sciences Michael Campbell David Machin amp Stephen Walters

Validity of a Survey

A survey is valid when it measures what it is supposed to measure

Content

Face

Criterion

Construct

19

Use of Correlation Coefficient for Reliability and Validity

Pearson correlation coefficients r are commonly used as measures of test-retest reliability inter-observer reliability and concurrent validity

Results could be misleading due to systematic biases

20

Determining the Sample Size

Survey should have enough power to find an effect if there exists one

Q What sample size would be enough to generalize results

Ans Determination of sample size depends upon study outcomes

Researcher must consider

Level of confidence

Margin of error 21

Survey Sampling Methods

Simple random sampling

Systematic random sampling

Stratified sampling

Cluster sampling

Convenience sampling

22

Analyzing Survey Data Descriptive Statistics

Measure of central tendency

Measure of dispersion

Inferential Statistics

Parametric amp non-parametric tests

Chi square

ANOVA

Regression amp correlation

More advanced techniques

23

Openstat A free software

Download

httpwwwstatprograms4ucomOpenStatMainhtm

Tutorial

httpmmc2geofisicaunammxcursosgeoestSoftwareOpenStatATextBookpdf

Youtube tutorials

24

Some Interesting Articles

Approaches to the Analysis of Survey Data httpwwwilriorgbiometricsTrainingResourcesDocumentsUniversity20of20ReadingGuides

Guides20on20AnalysisApprochAnalysispdf

Likert scales levels of measurement and the ldquolawsrdquo of statistics httpwwwfammedouhsceduresearchFMSRE20Orientation20amp20Handout20Materials

Handouts20520ScienceLikert20Scalespdf

Likert scales how to (ab)use them httpwwwgooglecomurlsa=tamprct=jampq=ampesrc=sampfrm=1ampsource=webampcd=1ampved=0CCoQFjAAampurl=http3A

2F2Fxayimgcom2Fkq2Fgroups2F187517252F1281694392Fname2F1likertscalespdfampei=LU8mUrGbI

MfD2QWe5YHwAQampusg=AFQjCNFQEMmMQbMzNc2go5XL-Cj-wo_rDgampsig2=HhUUy6KVsqqLqqXVxvLkdQampbvm=

bv51495398db2I

Essential elements of questionnaire design and development httpwwwbrighamandwomensorgmedical_professionalscareercfddMentoring20Resourcessurveydesignpdf

25

Thank you

QuestionsComments

RizwanaRehmanvagov

(919) 286-0411 ext 5024

26

For more information program materials and to complete evaluation for CME

credit visit wwwepilepsyvagovStatistics

Page 14: Statistics in Evidence Based Medicine II · Statistics in Evidence Based Medicine II ... Stages of survey research ... commonly used as measures of test-retest

Example of Test-retest Reliability

Juniper et al (1996) evaluated the pediatric Asthma Quality of Life Questionnaire (PAQLQ) by examining reliability in children aged 7 to 17 who had stable asthma

14

Item Within subject SD

Between subject SD

ICC

Overall QoL 017 073 095

Symptoms 022 084 093

Activities 042 096 084

Emotions 023 064 089

Medical Statistics A Text Book for the Health Sciences Michael Campbell David Machin amp Stephen Walters

Internal Consistency Reliability

A set of questions measure different aspects of one concept

Cronbachrsquos coefficient alpha (αChronbach) measures internal consistency reliability among a group of items to form a single scale

15

2

2

11 T

i

Chronbachs

s

k

k

k=number of items si=standard deviation of the ith item sT = standard deviation of the sum score T of all the items

Example of Internal Consistency

McKinley at al (1997) used a questionnaire to measure patient satisfaction with out-of ndashhours calls made to general practitioners They measured aspects such as satisfaction management and the doctorrsquos attitude They found values of αChronbach for each score ranging from 061 to 088 They concluded their questionnaire had satisfactory internal consistency

16

Medical Statistics A Text Book for the Health Sciences Michael Campbell David Machin amp Stephen Walters

Inter-observer(Inter-rater) Reliability

Inter-observer reliability provides a measure of how well two or more evaluators agree in their assessment of a variable A survey can be compared with a gold standard by using Cohenrsquos Kappa (k)

17

Chance

ChanceAgree

P

PP

1

PAgree= Estimated Probability the observerrsquos agree PChance= Proportion expected to show agreement by chance

Example of Inter-observer Reliability

18

Clamp and Kendrick (1998) used a telephone survey asking about safety to 165 families with children under age 5 years who had been involved in randomized trial concerning safety advice They chose a random sample of 20 families from the survey who then received a home visit 2 weeks later to measure the consistency of the response to the questions posed The investigators found that for most questions a high kappa value ( gt059) was obtained

Medical Statistics A Text Book for the Health Sciences Michael Campbell David Machin amp Stephen Walters

Validity of a Survey

A survey is valid when it measures what it is supposed to measure

Content

Face

Criterion

Construct

19

Use of Correlation Coefficient for Reliability and Validity

Pearson correlation coefficients r are commonly used as measures of test-retest reliability inter-observer reliability and concurrent validity

Results could be misleading due to systematic biases

20

Determining the Sample Size

Survey should have enough power to find an effect if there exists one

Q What sample size would be enough to generalize results

Ans Determination of sample size depends upon study outcomes

Researcher must consider

Level of confidence

Margin of error 21

Survey Sampling Methods

Simple random sampling

Systematic random sampling

Stratified sampling

Cluster sampling

Convenience sampling

22

Analyzing Survey Data Descriptive Statistics

Measure of central tendency

Measure of dispersion

Inferential Statistics

Parametric amp non-parametric tests

Chi square

ANOVA

Regression amp correlation

More advanced techniques

23

Openstat A free software

Download

httpwwwstatprograms4ucomOpenStatMainhtm

Tutorial

httpmmc2geofisicaunammxcursosgeoestSoftwareOpenStatATextBookpdf

Youtube tutorials

24

Some Interesting Articles

Approaches to the Analysis of Survey Data httpwwwilriorgbiometricsTrainingResourcesDocumentsUniversity20of20ReadingGuides

Guides20on20AnalysisApprochAnalysispdf

Likert scales levels of measurement and the ldquolawsrdquo of statistics httpwwwfammedouhsceduresearchFMSRE20Orientation20amp20Handout20Materials

Handouts20520ScienceLikert20Scalespdf

Likert scales how to (ab)use them httpwwwgooglecomurlsa=tamprct=jampq=ampesrc=sampfrm=1ampsource=webampcd=1ampved=0CCoQFjAAampurl=http3A

2F2Fxayimgcom2Fkq2Fgroups2F187517252F1281694392Fname2F1likertscalespdfampei=LU8mUrGbI

MfD2QWe5YHwAQampusg=AFQjCNFQEMmMQbMzNc2go5XL-Cj-wo_rDgampsig2=HhUUy6KVsqqLqqXVxvLkdQampbvm=

bv51495398db2I

Essential elements of questionnaire design and development httpwwwbrighamandwomensorgmedical_professionalscareercfddMentoring20Resourcessurveydesignpdf

25

Thank you

QuestionsComments

RizwanaRehmanvagov

(919) 286-0411 ext 5024

26

For more information program materials and to complete evaluation for CME

credit visit wwwepilepsyvagovStatistics

Page 15: Statistics in Evidence Based Medicine II · Statistics in Evidence Based Medicine II ... Stages of survey research ... commonly used as measures of test-retest

Internal Consistency Reliability

A set of questions measure different aspects of one concept

Cronbachrsquos coefficient alpha (αChronbach) measures internal consistency reliability among a group of items to form a single scale

15

2

2

11 T

i

Chronbachs

s

k

k

k=number of items si=standard deviation of the ith item sT = standard deviation of the sum score T of all the items

Example of Internal Consistency

McKinley at al (1997) used a questionnaire to measure patient satisfaction with out-of ndashhours calls made to general practitioners They measured aspects such as satisfaction management and the doctorrsquos attitude They found values of αChronbach for each score ranging from 061 to 088 They concluded their questionnaire had satisfactory internal consistency

16

Medical Statistics A Text Book for the Health Sciences Michael Campbell David Machin amp Stephen Walters

Inter-observer(Inter-rater) Reliability

Inter-observer reliability provides a measure of how well two or more evaluators agree in their assessment of a variable A survey can be compared with a gold standard by using Cohenrsquos Kappa (k)

17

Chance

ChanceAgree

P

PP

1

PAgree= Estimated Probability the observerrsquos agree PChance= Proportion expected to show agreement by chance

Example of Inter-observer Reliability

18

Clamp and Kendrick (1998) used a telephone survey asking about safety to 165 families with children under age 5 years who had been involved in randomized trial concerning safety advice They chose a random sample of 20 families from the survey who then received a home visit 2 weeks later to measure the consistency of the response to the questions posed The investigators found that for most questions a high kappa value ( gt059) was obtained

Medical Statistics A Text Book for the Health Sciences Michael Campbell David Machin amp Stephen Walters

Validity of a Survey

A survey is valid when it measures what it is supposed to measure

Content

Face

Criterion

Construct

19

Use of Correlation Coefficient for Reliability and Validity

Pearson correlation coefficients r are commonly used as measures of test-retest reliability inter-observer reliability and concurrent validity

Results could be misleading due to systematic biases

20

Determining the Sample Size

Survey should have enough power to find an effect if there exists one

Q What sample size would be enough to generalize results

Ans Determination of sample size depends upon study outcomes

Researcher must consider

Level of confidence

Margin of error 21

Survey Sampling Methods

Simple random sampling

Systematic random sampling

Stratified sampling

Cluster sampling

Convenience sampling

22

Analyzing Survey Data Descriptive Statistics

Measure of central tendency

Measure of dispersion

Inferential Statistics

Parametric amp non-parametric tests

Chi square

ANOVA

Regression amp correlation

More advanced techniques

23

Openstat A free software

Download

httpwwwstatprograms4ucomOpenStatMainhtm

Tutorial

httpmmc2geofisicaunammxcursosgeoestSoftwareOpenStatATextBookpdf

Youtube tutorials

24

Some Interesting Articles

Approaches to the Analysis of Survey Data httpwwwilriorgbiometricsTrainingResourcesDocumentsUniversity20of20ReadingGuides

Guides20on20AnalysisApprochAnalysispdf

Likert scales levels of measurement and the ldquolawsrdquo of statistics httpwwwfammedouhsceduresearchFMSRE20Orientation20amp20Handout20Materials

Handouts20520ScienceLikert20Scalespdf

Likert scales how to (ab)use them httpwwwgooglecomurlsa=tamprct=jampq=ampesrc=sampfrm=1ampsource=webampcd=1ampved=0CCoQFjAAampurl=http3A

2F2Fxayimgcom2Fkq2Fgroups2F187517252F1281694392Fname2F1likertscalespdfampei=LU8mUrGbI

MfD2QWe5YHwAQampusg=AFQjCNFQEMmMQbMzNc2go5XL-Cj-wo_rDgampsig2=HhUUy6KVsqqLqqXVxvLkdQampbvm=

bv51495398db2I

Essential elements of questionnaire design and development httpwwwbrighamandwomensorgmedical_professionalscareercfddMentoring20Resourcessurveydesignpdf

25

Thank you

QuestionsComments

RizwanaRehmanvagov

(919) 286-0411 ext 5024

26

For more information program materials and to complete evaluation for CME

credit visit wwwepilepsyvagovStatistics

Page 16: Statistics in Evidence Based Medicine II · Statistics in Evidence Based Medicine II ... Stages of survey research ... commonly used as measures of test-retest

Example of Internal Consistency

McKinley at al (1997) used a questionnaire to measure patient satisfaction with out-of ndashhours calls made to general practitioners They measured aspects such as satisfaction management and the doctorrsquos attitude They found values of αChronbach for each score ranging from 061 to 088 They concluded their questionnaire had satisfactory internal consistency

16

Medical Statistics A Text Book for the Health Sciences Michael Campbell David Machin amp Stephen Walters

Inter-observer(Inter-rater) Reliability

Inter-observer reliability provides a measure of how well two or more evaluators agree in their assessment of a variable A survey can be compared with a gold standard by using Cohenrsquos Kappa (k)

17

Chance

ChanceAgree

P

PP

1

PAgree= Estimated Probability the observerrsquos agree PChance= Proportion expected to show agreement by chance

Example of Inter-observer Reliability

18

Clamp and Kendrick (1998) used a telephone survey asking about safety to 165 families with children under age 5 years who had been involved in randomized trial concerning safety advice They chose a random sample of 20 families from the survey who then received a home visit 2 weeks later to measure the consistency of the response to the questions posed The investigators found that for most questions a high kappa value ( gt059) was obtained

Medical Statistics A Text Book for the Health Sciences Michael Campbell David Machin amp Stephen Walters

Validity of a Survey

A survey is valid when it measures what it is supposed to measure

Content

Face

Criterion

Construct

19

Use of Correlation Coefficient for Reliability and Validity

Pearson correlation coefficients r are commonly used as measures of test-retest reliability inter-observer reliability and concurrent validity

Results could be misleading due to systematic biases

20

Determining the Sample Size

Survey should have enough power to find an effect if there exists one

Q What sample size would be enough to generalize results

Ans Determination of sample size depends upon study outcomes

Researcher must consider

Level of confidence

Margin of error 21

Survey Sampling Methods

Simple random sampling

Systematic random sampling

Stratified sampling

Cluster sampling

Convenience sampling

22

Analyzing Survey Data Descriptive Statistics

Measure of central tendency

Measure of dispersion

Inferential Statistics

Parametric amp non-parametric tests

Chi square

ANOVA

Regression amp correlation

More advanced techniques

23

Openstat A free software

Download

httpwwwstatprograms4ucomOpenStatMainhtm

Tutorial

httpmmc2geofisicaunammxcursosgeoestSoftwareOpenStatATextBookpdf

Youtube tutorials

24

Some Interesting Articles

Approaches to the Analysis of Survey Data httpwwwilriorgbiometricsTrainingResourcesDocumentsUniversity20of20ReadingGuides

Guides20on20AnalysisApprochAnalysispdf

Likert scales levels of measurement and the ldquolawsrdquo of statistics httpwwwfammedouhsceduresearchFMSRE20Orientation20amp20Handout20Materials

Handouts20520ScienceLikert20Scalespdf

Likert scales how to (ab)use them httpwwwgooglecomurlsa=tamprct=jampq=ampesrc=sampfrm=1ampsource=webampcd=1ampved=0CCoQFjAAampurl=http3A

2F2Fxayimgcom2Fkq2Fgroups2F187517252F1281694392Fname2F1likertscalespdfampei=LU8mUrGbI

MfD2QWe5YHwAQampusg=AFQjCNFQEMmMQbMzNc2go5XL-Cj-wo_rDgampsig2=HhUUy6KVsqqLqqXVxvLkdQampbvm=

bv51495398db2I

Essential elements of questionnaire design and development httpwwwbrighamandwomensorgmedical_professionalscareercfddMentoring20Resourcessurveydesignpdf

25

Thank you

QuestionsComments

RizwanaRehmanvagov

(919) 286-0411 ext 5024

26

For more information program materials and to complete evaluation for CME

credit visit wwwepilepsyvagovStatistics

Page 17: Statistics in Evidence Based Medicine II · Statistics in Evidence Based Medicine II ... Stages of survey research ... commonly used as measures of test-retest

Inter-observer(Inter-rater) Reliability

Inter-observer reliability provides a measure of how well two or more evaluators agree in their assessment of a variable A survey can be compared with a gold standard by using Cohenrsquos Kappa (k)

17

Chance

ChanceAgree

P

PP

1

PAgree= Estimated Probability the observerrsquos agree PChance= Proportion expected to show agreement by chance

Example of Inter-observer Reliability

18

Clamp and Kendrick (1998) used a telephone survey asking about safety to 165 families with children under age 5 years who had been involved in randomized trial concerning safety advice They chose a random sample of 20 families from the survey who then received a home visit 2 weeks later to measure the consistency of the response to the questions posed The investigators found that for most questions a high kappa value ( gt059) was obtained

Medical Statistics A Text Book for the Health Sciences Michael Campbell David Machin amp Stephen Walters

Validity of a Survey

A survey is valid when it measures what it is supposed to measure

Content

Face

Criterion

Construct

19

Use of Correlation Coefficient for Reliability and Validity

Pearson correlation coefficients r are commonly used as measures of test-retest reliability inter-observer reliability and concurrent validity

Results could be misleading due to systematic biases

20

Determining the Sample Size

Survey should have enough power to find an effect if there exists one

Q What sample size would be enough to generalize results

Ans Determination of sample size depends upon study outcomes

Researcher must consider

Level of confidence

Margin of error 21

Survey Sampling Methods

Simple random sampling

Systematic random sampling

Stratified sampling

Cluster sampling

Convenience sampling

22

Analyzing Survey Data Descriptive Statistics

Measure of central tendency

Measure of dispersion

Inferential Statistics

Parametric amp non-parametric tests

Chi square

ANOVA

Regression amp correlation

More advanced techniques

23

Openstat A free software

Download

httpwwwstatprograms4ucomOpenStatMainhtm

Tutorial

httpmmc2geofisicaunammxcursosgeoestSoftwareOpenStatATextBookpdf

Youtube tutorials

24

Some Interesting Articles

Approaches to the Analysis of Survey Data httpwwwilriorgbiometricsTrainingResourcesDocumentsUniversity20of20ReadingGuides

Guides20on20AnalysisApprochAnalysispdf

Likert scales levels of measurement and the ldquolawsrdquo of statistics httpwwwfammedouhsceduresearchFMSRE20Orientation20amp20Handout20Materials

Handouts20520ScienceLikert20Scalespdf

Likert scales how to (ab)use them httpwwwgooglecomurlsa=tamprct=jampq=ampesrc=sampfrm=1ampsource=webampcd=1ampved=0CCoQFjAAampurl=http3A

2F2Fxayimgcom2Fkq2Fgroups2F187517252F1281694392Fname2F1likertscalespdfampei=LU8mUrGbI

MfD2QWe5YHwAQampusg=AFQjCNFQEMmMQbMzNc2go5XL-Cj-wo_rDgampsig2=HhUUy6KVsqqLqqXVxvLkdQampbvm=

bv51495398db2I

Essential elements of questionnaire design and development httpwwwbrighamandwomensorgmedical_professionalscareercfddMentoring20Resourcessurveydesignpdf

25

Thank you

QuestionsComments

RizwanaRehmanvagov

(919) 286-0411 ext 5024

26

For more information program materials and to complete evaluation for CME

credit visit wwwepilepsyvagovStatistics

Page 18: Statistics in Evidence Based Medicine II · Statistics in Evidence Based Medicine II ... Stages of survey research ... commonly used as measures of test-retest

Example of Inter-observer Reliability

18

Clamp and Kendrick (1998) used a telephone survey asking about safety to 165 families with children under age 5 years who had been involved in randomized trial concerning safety advice They chose a random sample of 20 families from the survey who then received a home visit 2 weeks later to measure the consistency of the response to the questions posed The investigators found that for most questions a high kappa value ( gt059) was obtained

Medical Statistics A Text Book for the Health Sciences Michael Campbell David Machin amp Stephen Walters

Validity of a Survey

A survey is valid when it measures what it is supposed to measure

Content

Face

Criterion

Construct

19

Use of Correlation Coefficient for Reliability and Validity

Pearson correlation coefficients r are commonly used as measures of test-retest reliability inter-observer reliability and concurrent validity

Results could be misleading due to systematic biases

20

Determining the Sample Size

Survey should have enough power to find an effect if there exists one

Q What sample size would be enough to generalize results

Ans Determination of sample size depends upon study outcomes

Researcher must consider

Level of confidence

Margin of error 21

Survey Sampling Methods

Simple random sampling

Systematic random sampling

Stratified sampling

Cluster sampling

Convenience sampling

22

Analyzing Survey Data Descriptive Statistics

Measure of central tendency

Measure of dispersion

Inferential Statistics

Parametric amp non-parametric tests

Chi square

ANOVA

Regression amp correlation

More advanced techniques

23

Openstat A free software

Download

httpwwwstatprograms4ucomOpenStatMainhtm

Tutorial

httpmmc2geofisicaunammxcursosgeoestSoftwareOpenStatATextBookpdf

Youtube tutorials

24

Some Interesting Articles

Approaches to the Analysis of Survey Data httpwwwilriorgbiometricsTrainingResourcesDocumentsUniversity20of20ReadingGuides

Guides20on20AnalysisApprochAnalysispdf

Likert scales levels of measurement and the ldquolawsrdquo of statistics httpwwwfammedouhsceduresearchFMSRE20Orientation20amp20Handout20Materials

Handouts20520ScienceLikert20Scalespdf

Likert scales how to (ab)use them httpwwwgooglecomurlsa=tamprct=jampq=ampesrc=sampfrm=1ampsource=webampcd=1ampved=0CCoQFjAAampurl=http3A

2F2Fxayimgcom2Fkq2Fgroups2F187517252F1281694392Fname2F1likertscalespdfampei=LU8mUrGbI

MfD2QWe5YHwAQampusg=AFQjCNFQEMmMQbMzNc2go5XL-Cj-wo_rDgampsig2=HhUUy6KVsqqLqqXVxvLkdQampbvm=

bv51495398db2I

Essential elements of questionnaire design and development httpwwwbrighamandwomensorgmedical_professionalscareercfddMentoring20Resourcessurveydesignpdf

25

Thank you

QuestionsComments

RizwanaRehmanvagov

(919) 286-0411 ext 5024

26

For more information program materials and to complete evaluation for CME

credit visit wwwepilepsyvagovStatistics

Page 19: Statistics in Evidence Based Medicine II · Statistics in Evidence Based Medicine II ... Stages of survey research ... commonly used as measures of test-retest

Validity of a Survey

A survey is valid when it measures what it is supposed to measure

Content

Face

Criterion

Construct

19

Use of Correlation Coefficient for Reliability and Validity

Pearson correlation coefficients r are commonly used as measures of test-retest reliability inter-observer reliability and concurrent validity

Results could be misleading due to systematic biases

20

Determining the Sample Size

Survey should have enough power to find an effect if there exists one

Q What sample size would be enough to generalize results

Ans Determination of sample size depends upon study outcomes

Researcher must consider

Level of confidence

Margin of error 21

Survey Sampling Methods

Simple random sampling

Systematic random sampling

Stratified sampling

Cluster sampling

Convenience sampling

22

Analyzing Survey Data Descriptive Statistics

Measure of central tendency

Measure of dispersion

Inferential Statistics

Parametric amp non-parametric tests

Chi square

ANOVA

Regression amp correlation

More advanced techniques

23

Openstat A free software

Download

httpwwwstatprograms4ucomOpenStatMainhtm

Tutorial

httpmmc2geofisicaunammxcursosgeoestSoftwareOpenStatATextBookpdf

Youtube tutorials

24

Some Interesting Articles

Approaches to the Analysis of Survey Data httpwwwilriorgbiometricsTrainingResourcesDocumentsUniversity20of20ReadingGuides

Guides20on20AnalysisApprochAnalysispdf

Likert scales levels of measurement and the ldquolawsrdquo of statistics httpwwwfammedouhsceduresearchFMSRE20Orientation20amp20Handout20Materials

Handouts20520ScienceLikert20Scalespdf

Likert scales how to (ab)use them httpwwwgooglecomurlsa=tamprct=jampq=ampesrc=sampfrm=1ampsource=webampcd=1ampved=0CCoQFjAAampurl=http3A

2F2Fxayimgcom2Fkq2Fgroups2F187517252F1281694392Fname2F1likertscalespdfampei=LU8mUrGbI

MfD2QWe5YHwAQampusg=AFQjCNFQEMmMQbMzNc2go5XL-Cj-wo_rDgampsig2=HhUUy6KVsqqLqqXVxvLkdQampbvm=

bv51495398db2I

Essential elements of questionnaire design and development httpwwwbrighamandwomensorgmedical_professionalscareercfddMentoring20Resourcessurveydesignpdf

25

Thank you

QuestionsComments

RizwanaRehmanvagov

(919) 286-0411 ext 5024

26

For more information program materials and to complete evaluation for CME

credit visit wwwepilepsyvagovStatistics

Page 20: Statistics in Evidence Based Medicine II · Statistics in Evidence Based Medicine II ... Stages of survey research ... commonly used as measures of test-retest

Use of Correlation Coefficient for Reliability and Validity

Pearson correlation coefficients r are commonly used as measures of test-retest reliability inter-observer reliability and concurrent validity

Results could be misleading due to systematic biases

20

Determining the Sample Size

Survey should have enough power to find an effect if there exists one

Q What sample size would be enough to generalize results

Ans Determination of sample size depends upon study outcomes

Researcher must consider

Level of confidence

Margin of error 21

Survey Sampling Methods

Simple random sampling

Systematic random sampling

Stratified sampling

Cluster sampling

Convenience sampling

22

Analyzing Survey Data Descriptive Statistics

Measure of central tendency

Measure of dispersion

Inferential Statistics

Parametric amp non-parametric tests

Chi square

ANOVA

Regression amp correlation

More advanced techniques

23

Openstat A free software

Download

httpwwwstatprograms4ucomOpenStatMainhtm

Tutorial

httpmmc2geofisicaunammxcursosgeoestSoftwareOpenStatATextBookpdf

Youtube tutorials

24

Some Interesting Articles

Approaches to the Analysis of Survey Data httpwwwilriorgbiometricsTrainingResourcesDocumentsUniversity20of20ReadingGuides

Guides20on20AnalysisApprochAnalysispdf

Likert scales levels of measurement and the ldquolawsrdquo of statistics httpwwwfammedouhsceduresearchFMSRE20Orientation20amp20Handout20Materials

Handouts20520ScienceLikert20Scalespdf

Likert scales how to (ab)use them httpwwwgooglecomurlsa=tamprct=jampq=ampesrc=sampfrm=1ampsource=webampcd=1ampved=0CCoQFjAAampurl=http3A

2F2Fxayimgcom2Fkq2Fgroups2F187517252F1281694392Fname2F1likertscalespdfampei=LU8mUrGbI

MfD2QWe5YHwAQampusg=AFQjCNFQEMmMQbMzNc2go5XL-Cj-wo_rDgampsig2=HhUUy6KVsqqLqqXVxvLkdQampbvm=

bv51495398db2I

Essential elements of questionnaire design and development httpwwwbrighamandwomensorgmedical_professionalscareercfddMentoring20Resourcessurveydesignpdf

25

Thank you

QuestionsComments

RizwanaRehmanvagov

(919) 286-0411 ext 5024

26

For more information program materials and to complete evaluation for CME

credit visit wwwepilepsyvagovStatistics

Page 21: Statistics in Evidence Based Medicine II · Statistics in Evidence Based Medicine II ... Stages of survey research ... commonly used as measures of test-retest

Determining the Sample Size

Survey should have enough power to find an effect if there exists one

Q What sample size would be enough to generalize results

Ans Determination of sample size depends upon study outcomes

Researcher must consider

Level of confidence

Margin of error 21

Survey Sampling Methods

Simple random sampling

Systematic random sampling

Stratified sampling

Cluster sampling

Convenience sampling

22

Analyzing Survey Data Descriptive Statistics

Measure of central tendency

Measure of dispersion

Inferential Statistics

Parametric amp non-parametric tests

Chi square

ANOVA

Regression amp correlation

More advanced techniques

23

Openstat A free software

Download

httpwwwstatprograms4ucomOpenStatMainhtm

Tutorial

httpmmc2geofisicaunammxcursosgeoestSoftwareOpenStatATextBookpdf

Youtube tutorials

24

Some Interesting Articles

Approaches to the Analysis of Survey Data httpwwwilriorgbiometricsTrainingResourcesDocumentsUniversity20of20ReadingGuides

Guides20on20AnalysisApprochAnalysispdf

Likert scales levels of measurement and the ldquolawsrdquo of statistics httpwwwfammedouhsceduresearchFMSRE20Orientation20amp20Handout20Materials

Handouts20520ScienceLikert20Scalespdf

Likert scales how to (ab)use them httpwwwgooglecomurlsa=tamprct=jampq=ampesrc=sampfrm=1ampsource=webampcd=1ampved=0CCoQFjAAampurl=http3A

2F2Fxayimgcom2Fkq2Fgroups2F187517252F1281694392Fname2F1likertscalespdfampei=LU8mUrGbI

MfD2QWe5YHwAQampusg=AFQjCNFQEMmMQbMzNc2go5XL-Cj-wo_rDgampsig2=HhUUy6KVsqqLqqXVxvLkdQampbvm=

bv51495398db2I

Essential elements of questionnaire design and development httpwwwbrighamandwomensorgmedical_professionalscareercfddMentoring20Resourcessurveydesignpdf

25

Thank you

QuestionsComments

RizwanaRehmanvagov

(919) 286-0411 ext 5024

26

For more information program materials and to complete evaluation for CME

credit visit wwwepilepsyvagovStatistics

Page 22: Statistics in Evidence Based Medicine II · Statistics in Evidence Based Medicine II ... Stages of survey research ... commonly used as measures of test-retest

Survey Sampling Methods

Simple random sampling

Systematic random sampling

Stratified sampling

Cluster sampling

Convenience sampling

22

Analyzing Survey Data Descriptive Statistics

Measure of central tendency

Measure of dispersion

Inferential Statistics

Parametric amp non-parametric tests

Chi square

ANOVA

Regression amp correlation

More advanced techniques

23

Openstat A free software

Download

httpwwwstatprograms4ucomOpenStatMainhtm

Tutorial

httpmmc2geofisicaunammxcursosgeoestSoftwareOpenStatATextBookpdf

Youtube tutorials

24

Some Interesting Articles

Approaches to the Analysis of Survey Data httpwwwilriorgbiometricsTrainingResourcesDocumentsUniversity20of20ReadingGuides

Guides20on20AnalysisApprochAnalysispdf

Likert scales levels of measurement and the ldquolawsrdquo of statistics httpwwwfammedouhsceduresearchFMSRE20Orientation20amp20Handout20Materials

Handouts20520ScienceLikert20Scalespdf

Likert scales how to (ab)use them httpwwwgooglecomurlsa=tamprct=jampq=ampesrc=sampfrm=1ampsource=webampcd=1ampved=0CCoQFjAAampurl=http3A

2F2Fxayimgcom2Fkq2Fgroups2F187517252F1281694392Fname2F1likertscalespdfampei=LU8mUrGbI

MfD2QWe5YHwAQampusg=AFQjCNFQEMmMQbMzNc2go5XL-Cj-wo_rDgampsig2=HhUUy6KVsqqLqqXVxvLkdQampbvm=

bv51495398db2I

Essential elements of questionnaire design and development httpwwwbrighamandwomensorgmedical_professionalscareercfddMentoring20Resourcessurveydesignpdf

25

Thank you

QuestionsComments

RizwanaRehmanvagov

(919) 286-0411 ext 5024

26

For more information program materials and to complete evaluation for CME

credit visit wwwepilepsyvagovStatistics

Page 23: Statistics in Evidence Based Medicine II · Statistics in Evidence Based Medicine II ... Stages of survey research ... commonly used as measures of test-retest

Analyzing Survey Data Descriptive Statistics

Measure of central tendency

Measure of dispersion

Inferential Statistics

Parametric amp non-parametric tests

Chi square

ANOVA

Regression amp correlation

More advanced techniques

23

Openstat A free software

Download

httpwwwstatprograms4ucomOpenStatMainhtm

Tutorial

httpmmc2geofisicaunammxcursosgeoestSoftwareOpenStatATextBookpdf

Youtube tutorials

24

Some Interesting Articles

Approaches to the Analysis of Survey Data httpwwwilriorgbiometricsTrainingResourcesDocumentsUniversity20of20ReadingGuides

Guides20on20AnalysisApprochAnalysispdf

Likert scales levels of measurement and the ldquolawsrdquo of statistics httpwwwfammedouhsceduresearchFMSRE20Orientation20amp20Handout20Materials

Handouts20520ScienceLikert20Scalespdf

Likert scales how to (ab)use them httpwwwgooglecomurlsa=tamprct=jampq=ampesrc=sampfrm=1ampsource=webampcd=1ampved=0CCoQFjAAampurl=http3A

2F2Fxayimgcom2Fkq2Fgroups2F187517252F1281694392Fname2F1likertscalespdfampei=LU8mUrGbI

MfD2QWe5YHwAQampusg=AFQjCNFQEMmMQbMzNc2go5XL-Cj-wo_rDgampsig2=HhUUy6KVsqqLqqXVxvLkdQampbvm=

bv51495398db2I

Essential elements of questionnaire design and development httpwwwbrighamandwomensorgmedical_professionalscareercfddMentoring20Resourcessurveydesignpdf

25

Thank you

QuestionsComments

RizwanaRehmanvagov

(919) 286-0411 ext 5024

26

For more information program materials and to complete evaluation for CME

credit visit wwwepilepsyvagovStatistics

Page 24: Statistics in Evidence Based Medicine II · Statistics in Evidence Based Medicine II ... Stages of survey research ... commonly used as measures of test-retest

Openstat A free software

Download

httpwwwstatprograms4ucomOpenStatMainhtm

Tutorial

httpmmc2geofisicaunammxcursosgeoestSoftwareOpenStatATextBookpdf

Youtube tutorials

24

Some Interesting Articles

Approaches to the Analysis of Survey Data httpwwwilriorgbiometricsTrainingResourcesDocumentsUniversity20of20ReadingGuides

Guides20on20AnalysisApprochAnalysispdf

Likert scales levels of measurement and the ldquolawsrdquo of statistics httpwwwfammedouhsceduresearchFMSRE20Orientation20amp20Handout20Materials

Handouts20520ScienceLikert20Scalespdf

Likert scales how to (ab)use them httpwwwgooglecomurlsa=tamprct=jampq=ampesrc=sampfrm=1ampsource=webampcd=1ampved=0CCoQFjAAampurl=http3A

2F2Fxayimgcom2Fkq2Fgroups2F187517252F1281694392Fname2F1likertscalespdfampei=LU8mUrGbI

MfD2QWe5YHwAQampusg=AFQjCNFQEMmMQbMzNc2go5XL-Cj-wo_rDgampsig2=HhUUy6KVsqqLqqXVxvLkdQampbvm=

bv51495398db2I

Essential elements of questionnaire design and development httpwwwbrighamandwomensorgmedical_professionalscareercfddMentoring20Resourcessurveydesignpdf

25

Thank you

QuestionsComments

RizwanaRehmanvagov

(919) 286-0411 ext 5024

26

For more information program materials and to complete evaluation for CME

credit visit wwwepilepsyvagovStatistics

Page 25: Statistics in Evidence Based Medicine II · Statistics in Evidence Based Medicine II ... Stages of survey research ... commonly used as measures of test-retest

Some Interesting Articles

Approaches to the Analysis of Survey Data httpwwwilriorgbiometricsTrainingResourcesDocumentsUniversity20of20ReadingGuides

Guides20on20AnalysisApprochAnalysispdf

Likert scales levels of measurement and the ldquolawsrdquo of statistics httpwwwfammedouhsceduresearchFMSRE20Orientation20amp20Handout20Materials

Handouts20520ScienceLikert20Scalespdf

Likert scales how to (ab)use them httpwwwgooglecomurlsa=tamprct=jampq=ampesrc=sampfrm=1ampsource=webampcd=1ampved=0CCoQFjAAampurl=http3A

2F2Fxayimgcom2Fkq2Fgroups2F187517252F1281694392Fname2F1likertscalespdfampei=LU8mUrGbI

MfD2QWe5YHwAQampusg=AFQjCNFQEMmMQbMzNc2go5XL-Cj-wo_rDgampsig2=HhUUy6KVsqqLqqXVxvLkdQampbvm=

bv51495398db2I

Essential elements of questionnaire design and development httpwwwbrighamandwomensorgmedical_professionalscareercfddMentoring20Resourcessurveydesignpdf

25

Thank you

QuestionsComments

RizwanaRehmanvagov

(919) 286-0411 ext 5024

26

For more information program materials and to complete evaluation for CME

credit visit wwwepilepsyvagovStatistics

Page 26: Statistics in Evidence Based Medicine II · Statistics in Evidence Based Medicine II ... Stages of survey research ... commonly used as measures of test-retest

Thank you

QuestionsComments

RizwanaRehmanvagov

(919) 286-0411 ext 5024

26

For more information program materials and to complete evaluation for CME

credit visit wwwepilepsyvagovStatistics