Difference between revisions of "Survey Pilot"

Jump to: navigation, search
(Created page with " = Guidelines for a Survey Pilot = == What is a survey pilot? == * A field test of the survey instrument and all survey protocols * Piloting is not just about the questionna...")
 
 
(223 intermediate revisions by 9 users not shown)
Line 1: Line 1:
 +
'''Survey pilot''' is the process of carrying out interviews and tests on different components of a survey, including [[Checklist: Content-focused Pilot|content]] and [[Checklist: Piloting Survey Protocols|protocols]]. A good '''pilot''' provides the research team with important feedback before they start the process of [[Primary Data Collection|primary data collection]]. This feedback can help the [[Impact Evaluation Team|research team]] review and improve the [[Questionnaire Design| design of the questionnaire]], [[Questionnaire Translation|translated questionnaires]], as well as [[Survey Protocols|survey protocols]] related to scheduling interviews, [[Sampling|sampling]], and [[Geo Spatial Data|geo data]].
 +
== Read First ==
 +
* A comprehensive '''survey pilot''' includes 3 stages: a pre-pilot, a content-focused pilot, and a data-focused pilot.
 +
* The '''pre-pilot''' is usually done during the process of [[Questionnaire Design|designing a questionnaire]], to refine the design, and wording of questions.
 +
* The '''content-focused pilot''' is conducted after [[Questionnaire Design|designing a questionnaire]], and sharing it with field and sector experts for their comments and inputs.
 +
* Finally, the '''data-focused pilot''' allows the [[Impact Evaluation Team|research team]] to review the data collected during the pilot, and identify possible errors in the [[Questionnaire Programming|programmed questionnaire]].
 +
* A '''survey pilot''' is different from the '''field testing''' (practice interviews) that enumerators conduct at the end of [[Enumerator Training|enumerator training]]. The pilot should be complete before enumerator training begins. 
 +
* [[Survey Pilot Participants|Participants of a survey pilot]] include the [[Impact Evaluation Team|research team]], '''interviewers''', and '''respondents'''.
 +
* Conduct the first two stages of the pilot before [[Questionnaire Programming|programming the questionnaire]]. Repeated programming is time-consuming and can create '''bugs''' (errors), for instance, by disturbing the order of questions.
  
= Guidelines for a Survey Pilot =
+
== Stages of a Survey Pilot ==
== What is a survey pilot? ==
+
A complete '''survey pilot''' is conducted over three stages - '''pre-pilot''', [[Checklist:_Content-focused_Pilot|content-focused pilot]], and [[Checklist:_Data-focused_Pilot|data-focused pilot]]. The table below discusses these stages in more detail.
  
* A field test of the survey instrument and all survey protocols
+
{| class="wikitable"
* Piloting is not just about the questionnaire: test all protocols and learn about relevant logistics
+
|-
* Piloting is done before Enumerator Training. It is not the same as the field practice all enumerators do at the end of Training!! It typically involves significant changes to the survey instrument and/or protocols – which should never be made after enumerator training!
+
! style="height:50px; width:150px; text-align:center;" |
* A complete survey pilot includes 3 stages (see Figure 1)
+
! style="height:50px; width:320px; text-align:center;" | Stage 1: Pre-pilot
 +
! style="height:50px; width:320px; text-align:center;" | Stage 2: Content-focused pilot
 +
! style="height:50px; width:320px; text-align:center;" | Stage 3: Data-focused pilot
 +
|-
 +
| style="text-align:center;" | '''Objective'''
 +
| style="vertical-align:top;"| Answer broad questions about <br>'''qualitative factors''' like [[Questionnaire Design|questionnaire design]] and the context for <br> conducting the study. Discuss these <br>with the concerned teams.
  
 +
Learn more about how people think to <br>see what structure and flow makes <br> sense.
 +
For example, do people think <br>about their input use at plot-level? <br>By crop? Overall?
  
* Are all 3 stages of piloting necessary for every survey? Not necessarily. If your survey is …
+
It helps to think of the '''pre-pilot''' as a part of the process of [[Questionnaire Design|questionnaire design]]. <br>For example, the '''pre-pilot''' can answer <br>specific questions such as - important sources of income, relevant types of shocks, local food groups, and so on.
** … a brand new survey instrument?  always start with Stage 1 – Pre-Pilot
 
** … an adaptation of a well-designed questionnaire from reliable source in the same country?
 
*** --> may start with Stage 2 – Content-focused pilot
 
** … an adaption of a survey instrument from a previous data collection for the same project?
 
*** If significant revisions or additions
 
**** -->  start with Stage 2
 
*** If no major changes
 
**** -->  skip to Stage 3 – Data -focused pilot
 
  
* Early stage pilots should be paper-based, even for CAPI surveys
+
In cases where the [[Impact Evaluation Team|research team]] is <br>using a pre-existing questionnaire, they <br>can skip this step. However, a '''pre-pilot''' <br>is absolutely essential if the '''research <br>team''' is designing a questionnaire from scratch, or if the questionnaire asks questions about issues that are <br>difficult to measure.
** Paper questionnaires are a more flexible way to record answers and qualitative observations
+
| style="vertical-align:top;"| Refine the order and wording of specific questions, the overall structure of the questionnaire, and [[Questionnaire Translation|translations]].
*** Record open-ended responses (critical for a pilot) more quickly / easily
+
 
*** Draw lines and arrows between questions to suggest restructuring
+
 
*** Record their observations and feedback in the margins
+
 
*** Make notes of questionnaire wording or translation problems directly in the text
+
Check that the answer [[SurveyCTO Choice Lists|choices]] are '''comprehensive''', that is, they cover all possibilities.
** A good pilot will provide significant inputs to questionnaire design, and result in significant changes to content and structure. Changing programming is time-consuming and can create bugs (e.g. if order of questions shifts and all skip codes need to be re-programmed)
+
 
 +
 
 +
 
 +
Flag any sensitive questions.
 +
 
 +
 
 +
 
 +
 
 +
 
 +
 
 +
Check how long it takes to answer the questions. Also check how answers differ between respondents ('''response variance''')
 +
| style="vertical-align:top;"| Check if the [[Questionnaire Programming|programmed instrument]] displays questions in the correct order, and follows the correct patterns (such as a group of questions which will always appear together, or will repeat).
 +
 
 +
Load a sample data set (based on '''pilot interviews'''). Check if the data set has any missing fields. Perform all data quality checks, like [[Back Checks|back checks]] and [[Monitoring Data Quality#High Frequency Checks|high frequency checks]].
 +
|-
 +
| style="height:80px;text-align:center;" | '''Status of survey instrument'''
 +
| style="vertical-align:center;" | Early, printable version of the draft, and notes for further discussion.
 +
| style="vertical-align:center;" | A translated, printable draft. Ready to be '''programmed'''.
 +
| style="vertical-align:center;" | A translated, programmed, final draft. Ready for starting data collection.
 +
|-
 +
| style="height:80px;text-align:center;" | '''Mode'''
 +
| style="text-align:center;"| Pen-and-paper
 +
| style="text-align:center;"| Pen-and-paper
 +
| style="text-align:center;"| Electronic (phone/tablet)
 +
|}
 +
 
 +
However, note that all three stages may not be necessary for every survey. The research team has to determine which stage to begin from on a case-to-case basis.
 +
* If the research team is using a brand new survey instrument, then they must start with Stage 1, the '''pre-pilot.'''
 +
* If the survey instrument is an adaptation of (based on) a different instrument which was used for a previous data collection for another project in the same region (province or country), the research team can start with Stage 2, the '''content-focused pilot'''. However, they must make sure that the survey instrument does not have any design issues and was shared by a reliable source.
 +
* If the survey instrument is an adaptation of an instrument used for a previous data collection for the same project, but the research team had to make significant revisions or additions, even then the research team should start with Stage 2, the '''content-focused pilot'''.
 +
* Only if the research team is '''piloting''' a '''follow-up''' survey, and there are no major changes from the '''baseline''' (or first round) survey, then in this case the research team may skip directly to Stage 3, the '''data-focused pilot'''.
 +
 
 +
== Pilot on paper first ==
 +
The first two stages of the survey pilot (pre-pilot and content-focused pilot) are best done on pen-and-paper is done on paper, regardless of the planned survey mode ([[Pen-and-Paper_Personal_Interviews_(PAPI)|pen-and-paper personal interviews (PAPI)]],
 +
[[Computer-Assisted_Personal_Interviews_(CAPI)|computer-assisted personal interviews (CAPI)]], [[Remote_Surveys|phone surveys (CATI)]]). A pen-and-paper pilot allows members of the research team and interviewers flexibility when it comes to recording answers and '''qualitative''' observations.
 +
 
 +
Piloting on paper is particularly helpful for the following:
 +
* '''Open-ended responses.''' pre-pilot responses help the research team to make the list of choices for a question more detailed and comprehensive. Pen-and-paper pilots allow the interviewers to make note of these.
 +
* '''Restructuring.''' Pen-and-paper pilots allow the interviewers to draw lines and arrows between questions, which makes it easier to restructure the '''instrument'''.
 +
* '''Observations.''' Pen-and-paper pilots allow interviewers to record observations and feedback of interviewers in the margins. This allows the research team to take note of issues, such as the wording or flow of questions.
 +
 
 +
== Timeline ==
 +
In order to conduct a successful [[Survey Pilot|survey pilot]], the [[Impact Evaluation Team|impact evaluation team]] must draft a detailed '''timeline for a survey pilot'''. The '''timeline''' should allow enough time for discussions and revisions about aspects like [[Questionnaire Design|instrument design]], [[Questionnaire Translation|translation]], and [[Survey Protocols|protocols]]. '''Piloting''' should ideally start 4-6 months before [[Field Surveys#Survey Launch|survey launch]] since this allows the research team to use the feedback from each [[Survey Pilot#Stages of a Survey Pilot|stage of a survey pilot]] to improve the survey content and [[Checklist: Piloting Survey Protocols|protocols]]. 
 +
 
 +
The '''timeline''' is an important part of the process of [[Structuring a Survey Pilot|structuring a survey pilot]]. The actual [[Structuring a Survey Pilot#Duration of the Pilot|duration of the pilot]] will depend on factors like number of changes made, location of the study area, and availability of equipment for '''field teams'''. Based on '''best practices''', the following are the steps involved in conducting a [[Survey Pilot|survey pilot]] (from start to finish), and the recommended time for each step.
 +
{| class="wikitable" style="margin-left: auto; margin-right: auto; border: none;"
 +
!style="width:350px; text-align:center;"| Task
 +
!style="width:350px; text-align:center;"| Estimated Time
 +
|-
 +
|style="text-align:center;"| [[Survey Pilot#Stages of a Survey Pilot|Pre-pilot]]
 +
|style="text-align:center;"| 1-2 weeks
 +
|-
 +
|style="text-align:center;"| [[Questionnaire Design|Questionnaire design]]
 +
|style="text-align:center;"| 4 weeks
 +
|-
 +
|style="text-align:center;"| [[Checklist: Content-focused Pilot|Content-focused pilot]]
 +
|style="text-align:center;"| 2-3 weeks
 +
|-
 +
|style="text-align:center;"| [[Structuring a Survey Pilot#Review Sessions|Review sessions]]
 +
|style="text-align:center;"| 1-2 weeks
 +
|-
 +
|style="text-align:center;"| [[Questionnaire Translation|Questionnaire translation]]
 +
|style="text-align:center;"| 2-3 weeks
 +
|-
 +
|style="text-align:center;"| [[IRB Approval|IRB approval]]
 +
|style="text-align:center;"| Depends on the complexity of the survey. Applying for [[IRB Approval|approvals]], [[Questionnaire Programming|programming]], and the [[Survey Pilot#Stages of a Survey Pilot|data-focused pilot]] - should all happen at the same time.
 +
|-
 +
|style="text-align:center;"| [[Questionnaire Programming|Questionnaire programming]]
 +
|style="text-align:center;"| 4-6 weeks
 +
|-
 +
|style="text-align:center;"| [[Survey Pilot#Stages of a Survey Pilot|Data-focused pilot]]
 +
|style="text-align:center;"| 2-3 weeks
 +
|-
 +
|style="text-align:center;"| [[Enumerator Training|Enumerator training]]
 +
|style="text-align:center;"| 2 weeks
 +
|}
 +
 
 +
Do not confuse the pilot with '''field testing''' (practice interviews) which is conducted at the end of [[Enumerator_Training |enumerator training]]!
 +
 
 +
== Structure ==
 +
The process of '''structuring a survey pilot''' involves agreeing upon the '''logistics''' (or practical aspects) of a [[Survey Pilot|survey pilot]], including '''duration''', '''approvals''', '''review sessions''', and '''training.''' In order to conduct a successful pilot, the [[Impact Evaluation Team|impact evaluation team]] (or research team) must discuss each of these aspects in detail to ensure that the '''survey pilot''' is conducted smoothly. A good pilot helps to improve the quality of the [[Primary Data Collection|data collection process]] and the [[Survey Protocols|survey protocols]]. Go to [[Structuring a Survey Pilot|Structuring a Survey Pilot]] for a detailed discussion of recommended protocols.
 +
 
 +
== Participants ==
 +
Typically, a '''survey pilot''' is conducted before a [[Survey Firm|survey firm]] is brought on-board. Each of the pilot participants plays an important role in the implementation of a successful pilot. Typical participants include respondents, interviewiers, field coordinator, principal investigator, and other research team members. For a detailed discussion of the role of each participant, and how to select appropriate respondents and interviewers for a pilot, go to the article on [[Survey Pilot Participants| Survey Pilot Participants]].
 +
 
 +
== Guidelines ==
 +
Finally, it is important to keep the following '''best practices''' in mind while planning and conducting a pilot:
 +
* Throughout the process of [[Questionnaire Design|designing the questionnaire]], take notes of what needs to be explored or tested during the pilot.
 +
* Encourage interviewers to probe and follow-up much more than they would in a typical interview
 +
* Encourage respondents to think out-loud, to understand how they are coming up with their answer, and ask enumerators to record these notes.
 +
* In some cultures, it will be appropriate and useful to ask respondents their feedback at the end of the interview
 +
* Take careful notes of the discussions and clarifications that come up during the pilot. These will be an important part of the [[Enumerator Training#Enumerator Manual|enumerator manual]] which is used during [[Enumerator Training|enumerator training]].
 +
* It is useful for research team members to sit in on pilot interviews even if they do not speak the local language. Take note of: questions that take longer than expected; any moments at which the respondent seems frustrated, confused, or uncomfortable; any questions that generate significant discussion; and the overall flow of modules and survey instrument.
 +
* Use the '''data-focused pilot''' to test [[Back Checks|back check]] templates.
 +
* Hire a local '''mobilizer''' to coordinate with respondents. '''Mobilizers''' explain the purpose behind conducting the survey, and facilitate the process of obtaining consent. This is particularly helpful in urban areas, or in cases where respondents are busy. This can improve the outcomes of piloting, for instance, by reducing gap between surveys.
 +
 
 +
[https://www.worldbank.org/en/research/dime/data-and-analytics DIME Analytics] has also created the following checklists to assist researchers and enumerators in preparing for, and implementing a pilot:
 +
*[[Preparing_for_the_survey_checklist|Checklist: Preparing for a survey pilot]]
 +
*[[Checklist: Content-focused Pilot|Checklist: Refining questionnaire content]]
 +
*[[Checklist: Data-focused Pilot|Checklist: Refining questionnaire data]]
 +
 
 +
== Related Pages ==
 +
[[Special:WhatLinksHere/Survey_Pilot|Click here for pages that link to this topic.]]
 +
 
 +
== Additional Resources ==
 +
* DIME Analytics (World Bank), [https://osf.io/63uv9/ Survey pilot guidelines]
 +
* DIME Analytics (World Bank), [https://osf.io/rqb5m/ Survey instrument design and pilot]
 +
[[Category: Research Design]]
 +
[[Category: Primary Data Collection]]

Latest revision as of 18:58, 18 May 2021

Survey pilot is the process of carrying out interviews and tests on different components of a survey, including content and protocols. A good pilot provides the research team with important feedback before they start the process of primary data collection. This feedback can help the research team review and improve the design of the questionnaire, translated questionnaires, as well as survey protocols related to scheduling interviews, sampling, and geo data.

Read First

  • A comprehensive survey pilot includes 3 stages: a pre-pilot, a content-focused pilot, and a data-focused pilot.
  • The pre-pilot is usually done during the process of designing a questionnaire, to refine the design, and wording of questions.
  • The content-focused pilot is conducted after designing a questionnaire, and sharing it with field and sector experts for their comments and inputs.
  • Finally, the data-focused pilot allows the research team to review the data collected during the pilot, and identify possible errors in the programmed questionnaire.
  • A survey pilot is different from the field testing (practice interviews) that enumerators conduct at the end of enumerator training. The pilot should be complete before enumerator training begins.
  • Participants of a survey pilot include the research team, interviewers, and respondents.
  • Conduct the first two stages of the pilot before programming the questionnaire. Repeated programming is time-consuming and can create bugs (errors), for instance, by disturbing the order of questions.

Stages of a Survey Pilot

A complete survey pilot is conducted over three stages - pre-pilot, content-focused pilot, and data-focused pilot. The table below discusses these stages in more detail.

Stage 1: Pre-pilot Stage 2: Content-focused pilot Stage 3: Data-focused pilot
Objective Answer broad questions about
qualitative factors like questionnaire design and the context for
conducting the study. Discuss these
with the concerned teams.

Learn more about how people think to
see what structure and flow makes
sense. For example, do people think
about their input use at plot-level?
By crop? Overall?

It helps to think of the pre-pilot as a part of the process of questionnaire design.
For example, the pre-pilot can answer
specific questions such as - important sources of income, relevant types of shocks, local food groups, and so on.

In cases where the research team is
using a pre-existing questionnaire, they
can skip this step. However, a pre-pilot
is absolutely essential if the research
team
is designing a questionnaire from scratch, or if the questionnaire asks questions about issues that are
difficult to measure.

Refine the order and wording of specific questions, the overall structure of the questionnaire, and translations.


Check that the answer choices are comprehensive, that is, they cover all possibilities.


Flag any sensitive questions.




Check how long it takes to answer the questions. Also check how answers differ between respondents (response variance)

Check if the programmed instrument displays questions in the correct order, and follows the correct patterns (such as a group of questions which will always appear together, or will repeat).

Load a sample data set (based on pilot interviews). Check if the data set has any missing fields. Perform all data quality checks, like back checks and high frequency checks.

Status of survey instrument Early, printable version of the draft, and notes for further discussion. A translated, printable draft. Ready to be programmed. A translated, programmed, final draft. Ready for starting data collection.
Mode Pen-and-paper Pen-and-paper Electronic (phone/tablet)

However, note that all three stages may not be necessary for every survey. The research team has to determine which stage to begin from on a case-to-case basis.

  • If the research team is using a brand new survey instrument, then they must start with Stage 1, the pre-pilot.
  • If the survey instrument is an adaptation of (based on) a different instrument which was used for a previous data collection for another project in the same region (province or country), the research team can start with Stage 2, the content-focused pilot. However, they must make sure that the survey instrument does not have any design issues and was shared by a reliable source.
  • If the survey instrument is an adaptation of an instrument used for a previous data collection for the same project, but the research team had to make significant revisions or additions, even then the research team should start with Stage 2, the content-focused pilot.
  • Only if the research team is piloting a follow-up survey, and there are no major changes from the baseline (or first round) survey, then in this case the research team may skip directly to Stage 3, the data-focused pilot.

Pilot on paper first

The first two stages of the survey pilot (pre-pilot and content-focused pilot) are best done on pen-and-paper is done on paper, regardless of the planned survey mode (pen-and-paper personal interviews (PAPI), computer-assisted personal interviews (CAPI), phone surveys (CATI)). A pen-and-paper pilot allows members of the research team and interviewers flexibility when it comes to recording answers and qualitative observations.

Piloting on paper is particularly helpful for the following:

  • Open-ended responses. pre-pilot responses help the research team to make the list of choices for a question more detailed and comprehensive. Pen-and-paper pilots allow the interviewers to make note of these.
  • Restructuring. Pen-and-paper pilots allow the interviewers to draw lines and arrows between questions, which makes it easier to restructure the instrument.
  • Observations. Pen-and-paper pilots allow interviewers to record observations and feedback of interviewers in the margins. This allows the research team to take note of issues, such as the wording or flow of questions.

Timeline

In order to conduct a successful survey pilot, the impact evaluation team must draft a detailed timeline for a survey pilot. The timeline should allow enough time for discussions and revisions about aspects like instrument design, translation, and protocols. Piloting should ideally start 4-6 months before survey launch since this allows the research team to use the feedback from each stage of a survey pilot to improve the survey content and protocols.

The timeline is an important part of the process of structuring a survey pilot. The actual duration of the pilot will depend on factors like number of changes made, location of the study area, and availability of equipment for field teams. Based on best practices, the following are the steps involved in conducting a survey pilot (from start to finish), and the recommended time for each step.

Task Estimated Time
Pre-pilot 1-2 weeks
Questionnaire design 4 weeks
Content-focused pilot 2-3 weeks
Review sessions 1-2 weeks
Questionnaire translation 2-3 weeks
IRB approval Depends on the complexity of the survey. Applying for approvals, programming, and the data-focused pilot - should all happen at the same time.
Questionnaire programming 4-6 weeks
Data-focused pilot 2-3 weeks
Enumerator training 2 weeks

Do not confuse the pilot with field testing (practice interviews) which is conducted at the end of enumerator training!

Structure

The process of structuring a survey pilot involves agreeing upon the logistics (or practical aspects) of a survey pilot, including duration, approvals, review sessions, and training. In order to conduct a successful pilot, the impact evaluation team (or research team) must discuss each of these aspects in detail to ensure that the survey pilot is conducted smoothly. A good pilot helps to improve the quality of the data collection process and the survey protocols. Go to Structuring a Survey Pilot for a detailed discussion of recommended protocols.

Participants

Typically, a survey pilot is conducted before a survey firm is brought on-board. Each of the pilot participants plays an important role in the implementation of a successful pilot. Typical participants include respondents, interviewiers, field coordinator, principal investigator, and other research team members. For a detailed discussion of the role of each participant, and how to select appropriate respondents and interviewers for a pilot, go to the article on Survey Pilot Participants.

Guidelines

Finally, it is important to keep the following best practices in mind while planning and conducting a pilot:

  • Throughout the process of designing the questionnaire, take notes of what needs to be explored or tested during the pilot.
  • Encourage interviewers to probe and follow-up much more than they would in a typical interview
  • Encourage respondents to think out-loud, to understand how they are coming up with their answer, and ask enumerators to record these notes.
  • In some cultures, it will be appropriate and useful to ask respondents their feedback at the end of the interview
  • Take careful notes of the discussions and clarifications that come up during the pilot. These will be an important part of the enumerator manual which is used during enumerator training.
  • It is useful for research team members to sit in on pilot interviews even if they do not speak the local language. Take note of: questions that take longer than expected; any moments at which the respondent seems frustrated, confused, or uncomfortable; any questions that generate significant discussion; and the overall flow of modules and survey instrument.
  • Use the data-focused pilot to test back check templates.
  • Hire a local mobilizer to coordinate with respondents. Mobilizers explain the purpose behind conducting the survey, and facilitate the process of obtaining consent. This is particularly helpful in urban areas, or in cases where respondents are busy. This can improve the outcomes of piloting, for instance, by reducing gap between surveys.

DIME Analytics has also created the following checklists to assist researchers and enumerators in preparing for, and implementing a pilot:

Related Pages

Click here for pages that link to this topic.

Additional Resources