CAPACITY BUILDING & TRAINING

We assist national statistics offices and survey organizations worldwide with the development of their technical skills

CAPACITY BUILDING & TRAINING

We assist national statistics offices and survey organizations worldwide with the development of their technical skills

For over thirty years Sistemas Integrales has assisted international organizations and statistics offices worldwide with the development of their technical capabilities through training on all elements of survey methods. Our courses and workshops cover both theoretical and practical elements of the survey practice and are delivered by international experts. They allow participants to examine firsthand experiences in concrete surveys, including problems encountered and solutions adopted. All meetings can be conducted in the client´s country in English, French, Spanish or Portuguese.

Workshop for Designing and Implementing Phone-based Surveys

 

This is a hands-on workshop that covers all the design and implementation topics necessary to conduct a phone-based survey, either as a follow-up to a face-to-face survey or as a stand-alone phone survey. Participants will review all theoretical and practical aspects related to sampling and estimation for a phone survey, as well as the stages and details involved in the operations. They will receive expert step-by-step guidance on logistics, infrastructure, questionnaire design, sample management, interviewer training, fieldwork organization, data management and quality control.

The concepts learned and the practice gained in the workshop will allow participants to design and implement a phone survey from scratch or shift a running face-to-face survey to phone when circumstances call for it. All topics will be treated in multiple practice sessions based on the specific phone surveys to be implemented by the participants.

Duration: 5 full days or 10 half days.

Prerequisites: Experience in the management and execution of household surveys. Working knowledge of Excel and Stata or SPSS.

Requirements: Each participant should work with Stata or SPSS and a visualization application on his/her own computer.

CONTENTS

MODULE 1. SAMPLE DESIGN
1. Two scenarios: A) follow-up of in-person surveys, B) RDD sampling (Random Digit Dialing)
2. Dual frames. Mobile phone multiplicity adjustment.
3. Sampling frame coverage: landline and mobile phone coverage. Potential coverage bias.
4. Nonresponse (refusals + no-contacts). High nonresponse levels. Potential nonresponse bias.

MODULE2. WEIGHTING AND ESTIMATION
1. RDD Phone sample design and selection. Sample replicates.
2. Weighting with dual frames and multiplicity adjustment.
3. Nonresponse adjustment under scenarios A and B. Response propensity models. Calibration (poststratification and raking).
4. Sampling error estimation.
5. Comparability with face-to-face survey estimates. Mode effects. Differential coverage and response rates. One-shot measurements and time series.

MODULE3. IMPLEMENTATION AND DATA MANAGEMENT
1. Questionnaire design in phone surveys.
2. Central location and decentralization.
3. CATI (Computer Assisted Telephone Interviewing). On-site and remote CATI systems. Functionalities. VOIP (Voice on Internet Protocol).
4. Quick solutions for short-term projects. CAPI (Computer Assisted Personal Interviewing) software using tablets and mobile phones.
5. Sample management. Replicates and batches.
6. Call protocols and strategies for reducing nonresponse.
7. Interviewer training for phone surveys.
8. Mixed-mode data collection. Mixed modes. Responsive Survey Design. What is Responsive Survey Design?

MODULE 4. QUALITY CONTROL
1. Sampling and nonsampling errors. The Total Survey Error Framework. Biases and variances.
2. Potential coverage bias and potential nonresponse bias. Determinants of nonresponse and strategies for reducing unit and item nonresponse.
3. Measurement Error. Sources of measurement error. Interviewer design effect. Advantages of an interpenetrating design.
4. Innovative and traditional Quality Control tools. Intra-questionnaire checks (CAPI and CAFE). Types of intra-questionnaire checks. Batch controls.
5. Monitoring of aggregate quality indicators. Types of indicators based on data and paradata. The Interviewer Risk Index. Disposition codes and outcome rates. Identification of meaningful Quality Indicators. Dashboards and data visualization software.
6. Back-checks, randomized audio supervision (RAS), GPS, mystery respondents. The importance of training. Why not data cleaning?
7. Training of interviewers and supervisors
8. Fieldwork organization. Staff roles, responsibilities, payment schemes. Effective feedback cycles for field teams.
9. Setup of a Comprehensive Survey Quality Control System (CSQC).

 

Workshop on Survey Quality Control and Data Management

 

International survey practice usually gives a great deal of attention to minimizing sampling error, whereas efforts to keep nonsampling error under control are often neglected. However, nonsampling errors may have severe implications for the accuracy of the data collected, and standard data cleaning at the end of fieldwork is not a solution.

This is a hands-on workshop that covers complementary strategies for detecting and reducing different types of survey errors generated during the field stage, such as coverage error, nonresponse error and interviewer effects on measurement error. The workshop starts by presenting the Total Survey Error framework, which identifies all the possible sources of error present in every survey. It then introduces a set of techniques and procedures that are implemented throughout the full fieldwork period to detect errors in a timely fashion and take immediate action while still in the field. Such techniques include intra-questionnaire checks, monitoring of aggregate quality indicators, back-checks and randomized audio supervision, plus suitable corrective interventions for dealing with the identified problems. Finally, the workshop will offer detailed guidelines for setting up a Comprehensive Survey Quality Control system.

Classes will provide numerous examples based on recent surveys conducted in different countries and on diverse subjects such as labor, household budget, poverty, health, nutrition, gender violence and agriculture. The workshop will include numerous practice sessions where participants will have the opportunity to work with the instruments and data from their own surveys and apply the quality control methods learned. The class will use Stata or SPSS and a data visualization application. 

By the end of the workshop, participants will:

  • be able to recognize the various sources of error in every survey
  • know how to implement a range of techniques to detect different types of error produced in the field of their own surveys
  • know what actions can be taken to prevent and reduce those errors, thus improving the accuracy and reliability of the data they produce
  • be able to assess the quality of any survey data before its analysis
  • know how to set up a Comprehensive Survey Quality Control system for their own surveys

Duration: 5 full days or 10 half days.

Prerequisites: Experience in the management and execution of household surveys. Working knowledge of Excel and Stata or SPSS.

Requirements: Each participant should work with Stata or SPSS and a visualization application on his/her own computer.

CONTENTS

MODULE 1
1. Why Survey Quality Control? Quality assurance and quality control. Sampling and nonsampling errors. The Total Survey Error Framework. Biases and variances.
2. Coverage error and listing operations. Coverage bias. Interviewer effects.
3. Unit nonresponse and item nonresponse. Nonresponse bias. Interviewer effects. Determinants of nonresponse and strategies for reducing unit and item nonresponse.
4. Measurement Error. Sources of measurement error. Interviewer bias and interviewer variance. Intra-interviewer correlation. Interviewer design effect. Then need for interpenetrating design.

MODULE 2
1. Innovative and traditional Quality Control tools.
2. Intra-questionnaire checks (CAPI and CAFE). Types of intra-questionnaire checks.
3. Batch controls.
4. Monitoring of aggregate quality indicators. Types of indicators based on data and paradata.
5. The Interviewer Risk Index.
6. Disposition codes and outcome rates.
7. Identification and computation of meaningful Quality Indicators.
8. Display of Quality Indicators. Introduction to data visualization software.
9. Back-checks, randomized audio supervision (RAS), GPS, mystery respondents. The importance of training. Why not data cleaning?

MODULE 3
1. Training of interviewers and supervisors.
2. Fieldwork organization. Staff roles, responsibilities, payment schemes. Effective feedback cycles for field teams.
3. Setup of a Comprehensive Survey Quality Control System.

Applied Sampling I: Survey Sampling and Weighting

 

This course covers sampling theory and practice of sample design, weighting and estimation, with an emphasis on applications. Examples are based on actual surveys, and in-class exercises will give participants hands-on experience with the techniques learned. Participants will also gain an understanding of topics and challenges common to the survey sampling practice. The course is aimed at managers and technical staff working within governments or international agencies and involved in programs on poverty, employment, health, education and living conditions. The course can be delivered in English, French, Spanish and Portuguese.

Duration: 10 full days or 20 half days.

Prerequisites: Notions of probability theory and mathematical statistics. Working knowledge of Excel and Stata or SPSS.

Requirements: Each participant should work with Excel and Stata on his/her own computer.

CONTENTS

1. Probability sampling and why sampling needs to be random. Types of non-probability samples and their pitfalls. Population and sampling distributions. Notions of estimator, estimate and sampling error. The Central Limit Theorem. Unbiased estimators, precision and accuracy.

2. Simple Random Sampling and Systematic sampling. Estimation of a proportion. Sampling error: sampling variance, standard error, confidence intervals and coefficient of variation. How are sampling errors affected by the population size, the sample size, and the estimated parameter. Estimation of a mean.

3. Exercises with Simple Random Sampling and Systematic sampling: sample selection and estimation.

4. Stratified sampling. Strata and domains. Statistical efficiency of a sample design. Advantages and disadvantages of different allocation criteria. Sample allocation simulations based on census data. Exercises.

5. Cluster sampling and two-stage sampling. Advantages and downsides. Intraclass correlation and design effects. Area sampling. Unequal sized cluster sampling. Selection of primary sampling units with probabilities proportionate to size and selection of secondary sampling units with equal probabilities. Exercises.

6. Complex area sampling. First- and second-stage sample frames. Segmenting. Stratification and selection of areas with probability proportionate to size using census data. Household listing operations. Selection of dwellings and households. Synthetic design effects. Exercises using data from listing operations.

7. Weighting techniques. Base weights, nonresponse adjustment and calibration (poststratification and raking). Exercises on the computation of weights under different sample designs.

8. Sampling error estimation for complex samples using Stata. The ultimate cluster approach. Linearized variance estimation. Exercises and examples using actual survey data.

Final examination.

INTRODUCTIoN TO THE aNALYSIS OF SURVEY DATa

This is a hands-on workshop whose ultimate objective is that participants gain the skills necessary to basic statistical analysis based on data from various types of surveys, such ar labor force, poverty, agricultural and living conditions surveys. The workshop presents how data is typically organized, introduces key descriptive statistics concepts, and finally addresses the formulation, computation and interpretation of indicators for different policy areas. Conceptual topics will be followed by applications on data from actual surveys using the Stata statistical software. Classes will include practice sessions and a capstone exercise where participants will have the opportunity to apply the methods learned using data from real surveys. The workshop can be delivered in English, French, Spanish and Portuguese.

Duration: 8 sessions of 3 hours.

Prerequisites: none.

Requirements: Each participant should work with Stata on his/her own computer.

CONTENTS
1. OVERVIEW OF SURVEY STATISTICS
1. Household, population and agricultural data sources. Their characteristics, advantages and limitations.
2. Differences among data, information and knowledge.
Discussion.

2. MICRODATA ORGANIZATION
1. The dataset: observations or analysis units, variables and values.
2. Data codebooks.
3. Overview of the Stata statistical software: screens and toolbar, using and saving data, do files, log files.
4. Most frequent Stata commands.

3. GENERATING NEW VARIABLES
1. Numeric and categorical variables. Measurement levels.
2. Generation of new variables with Stata.
3. Recoding of existing variables with Stata.
Examples on generating and transforming variables based on survey data.

4. DESCRIPTIVE STATISTICS
1. Categorical variables: proportions and percentages.
2. Univariate and multivariate distributions, contingency tables. Graphics.
3. Numeric variables: mean, median and quantiles, variance and standard deviation. Graphics.
4. Stata commands and applications.
5. Interpretation and analysis.
Exercises based on household and agricultural survey data.

5. FORMULATING, COMPUTING AND ANALYZING POLICY-RELEVANT INDICATORS
1. Definition and utility of an indicator.
2. Types of indicators: proportion, rate, ratio, index.
3. The five features of a good indicator.
4. The five stages of their generation process.
5. Frequent problems found in their generation.
6. Relevant breakdown variables.
7. Combination of data sources.
8. Indicator technical summary files.
Applications covering the entire process for formulating and calculating selected indicators based on data from different surveys.

6. BASIC SAMPLING-RELATED CONCEPTS
1. Why does sample selection need to be random? Estimation bias.
2. What is probability sampling? And what is it not? Representation.
3. Most common sample designs in agricultural surveys.
4. Missing data: unit and item nonresponse. How can it be handled?
5. The use of sample weights when working with survey data.

7. ESTIMATION BASED ON COMLPEX-SAMPLE SURVEY DATA
1. Use of the svy commands to account for the sample design (stratification and clustering) and weighting.
2. Measures of precision of the computed survey estimates: standard errors, confidence intervals and coefficients of variation.
3. When can a survey estimate be considered good/precise enough?
Applications

8. CAPSTONE EXERCISE ON THE GENERATION AND ANALYSIS OF KEY INDICATORS

 

workshop on survey methods

 

Drawing on our extensive international experience, we train multilateral and government agencies in survey methodology to help them develop the skills required to execute and manage a specific survey. Over the years, we have delivered a large number of hands-on workshops to assist with surveys of all types, including studies on living standards, labor force, income and expenditure, health and education, among others. This training typically takes place at the statistics bureau of the country where the survey is conducted or at our premises.

This workshop has a learning-by-doing approach and covers all practical aspects of sample design, instrument conception, fieldwork organization, quality control and budgeting, which are all key elements for producing valid, reliable and consistent data. All sessions are delivered by international experts.

This workshop is aimed at professionals from government and multilateral agencies, statistics offices and academics. It is intended for technical agents who work in the management and implementation of surveys. It is also ideal for people responsible for procuring survey services or who use survey data for the formulation, monitoring and evaluation of public policies in areas such as employment, health, education, poverty, nutrition, living conditions, etc. It can be delivered in all countries, in English, French, Spanish and Portuguese. 

Objectives

  • Understand the importance of producing and analyzing valid and reliable data.
  • Identify and analyze the stages of the survey cycle, covering both design and implementation.
  • Gain the skills to develop and manage a survey project: sample design, instrument conception, fieldwork organization, quality control and budgeting.
  • Gain the skills to produce thorough and precise terms of reference for survey projects to be implemented by third parties.
  • Gain the necessary understanding to assess proposals submitted by survey firms.

Address

Jose M de la Barra 412, piso 4
Postcode 6500446
Santiago, Chile

Phone

+562 2638 1841
+562 2639 4554

Sistemas Integrales 2018 - All Rights Reserved

Address

José M de la Barra 412, piso 4
Postcode 6500446
Santiago, Chile

Phone

+562 2638 1841
+562 2639 4554

Sistemas Integrales 2018 – All Rights Reserved