2. When It Comes to Surveys, Much to Complain About
Too time consuming
Ask too many personal questions
Come at inconvenient times
Are too long
Can't be trusted
0% 5% 10% 15% 20% 25%
Source: Vovici, n = 100
2
3. What Makes a Successful Survey?
Meets its original goal
Produces accurate data that can be used
to make decisions from
Improves the satisfaction level of its
respondents
Used to improve the sponsoring
organization
3
4. The Seven Steps of Highly Successful Surveyors
1. Focus on a Goal
2. Survey the Right Group of People
3. Craft Your Invitation Carefully
4. Order Questions Logically
5. Write Objective Questions
6. Shorten the Survey
7. Close the Feedback Loop
4
5. Agenda
Response Rate Targets
Drivers of Response Rates
Jumping the Hurdles to Response
Craft Your Invitation Carefully
Reporting
Charts
Q&A
Multi-Language Surveys
5
7. Oversimplified Look at Market Research “Rates”
General Population
Invitees
Incidence Rate = Invitees ÷ General Population
Invitation Readers
Participants
Participation Rate = Participants ÷ Invitees Click-Thru Rate = Participants ÷ Invitation Readers
Completes
Response Rate 1 ~= Completes ÷ Invitees Completion Rate = Completes ÷ Participants
7
8. Response Rate Goals
Compliance
Representative at Higher Confidence
Representative by Cell
Representative Overall
8
9. High Response Rates Not Needed
Higher response rates do not produce much more accurate results
Fear is that low response rates indicate a potential non-response bias
Pew Research Center found a 25% response rate did not reduce quality of
survey estimates compared to a 50% response rate
Source: “Gauging the Impact of Growing Nonresponse on Estimates from a National RDD Telephone Survey”
Holbrook, Krosnick and Pfent found:
Little variance by response rate
But aim for response rates of at least 20% (“Our findings might not generalize ... to
declines in response rates below 20 percent.”)
Source: “The Causes and Consequences of Response Rates in Surveys by the News Media and Government
Contractor Survey Research Firms”
11. Key Drivers to Response Rates
Target
Quantity of Audience
Reminders
History of
Invitation
Frequency
Empanelment
Invitation
Salience
Recency
Exclusivity
Incentives
11
12. Target Audience Matters
Response Rates Proportional to Strength of Relationship
100%
80%
60%
40%
20% Low
High
0%
Average
Source: Vovici
12
13. Use Reminders to Improve Response Rate
Cumulative Response Rate
35%
30%
25%
20%
15% Optimize rather than
maximize response rates
10%
5%
0%
Invitation 1st Reminder 2nd Reminder 3rd Reminder
13
14. History of Invitation Frequency
For customers who aren’t members of a
panel:
The Boy Who Cried “Survey!” – too
frequent requests for feedback lead to a
decline in response rates
Rare requests to participate in a survey
produce a higher response
Response rates suffer if Marketing sends
out a high volume of email (“spam”)
14
15. Empanelment
Respondents who have opted in to a panel
expect to participate
Engaged panelists will complain if they are
not surveyed each month
15
16. Invitation Salience: Effect of Topic & Text on Response
• Participate and earn a free gift card from Home Depot
• Participate and win one of 3 cash prizes
• Participate and $5 will be credited to the credit card of your choice
• We’d like to know your thoughts about gasoline prices •This survey is super-short! (we promise)
• Shorter than the typical survey! Takes only 10 minutes… • We only need to ask you about 10 questions
• Your answers will remain completely confidential • We only need to ask you about 25 questions
• We’d like to ask you questions about home improvement
• We’d like to ask you questions about Space Tourism
• Get a free copy of the results after you participate!
• The results of the survey will be published in the New York Times!
• Earn frequent flyer miles on the credit card of your choice
Source: CMOR, “Creating a Respondent Cooperation Playbook”
16
17. Recency
The longer the lag between the event and
the survey the lower the response rate.
The response rate for one transactional
survey jumped from 15% to 63% by
sending invite within 15 minutes of event
instead of 24 hours later.
17
18. Exclusivity
Emphasizing the exclusivity of the
invitation will improve response rates.
Example: "You are one of a small, select
group of customers that we have invited to
provide us feedback."
Of course, don't say it if it isn't true!
18
19. Incentives & Response
“It has been repeatedly shown that providing financial
incentives as a persuasive motivator for reluctant
respondents is a viable and effective technique for
increasing response rates…” - “Demographics and Online
Survey Response Rates”, Kurt Knapton and Steve Myers
“....Goetz, Tyler and Cook…concluded that financial
incentives increased response rates and showed no
differences in the demographics of the incentive and
non-incentive groups.”
“....Mason, Lesser and Dillman (2003) found that non-
response error may actually be reduced when financial
incentives are used...”
19
20. Meta Analysis
“Factors affecting response rates of the
web survey: A systematic review”
Computers in Human Behavior (March 2010)
Weimiao Fan and Zheng Yan
Synthesizes results of 300 studies into a
model of the contributing factors of
response rates
Response rate = completes / invites
20
21. Factors Affecting Response Rates to Web Surveys
Content of
Target Invitation Technology
web Survey delivery
audience designs issues
questionnaires
Nature of the
Type of sponsors of the Sampling Personalization
population survey methods
Survey software
What the topic is Scarcity
Contact delivery
Demographics
modes
How long the
survey takes to Access control
complete Internet
transmission
Personality Contact quality capabilities
Question writing Use of incentives
Source: Fan and Yan, 2010
21
22. The Science of Email Marketing – Applied to Surveys
Businesses are consumers.
Try sending survey invitations on the weekends.
Send invites very early in the morning.
Use lots of links in your survey invites.
Include reference information in your survey invitation, providing recipients
other ways to provide feedback or seek assistance after the survey is closed.
Subject lines should use the word survey and should not emphasize incentives.
Emphasize in the invite that this is an exclusive opportunity to provide
feedback (for those studies for which it is true, of course).
Source: Dan Zarrella, “The Science of Email Marketing”, 2011
22
24. The Third Step: Craft Your Invitation Carefully
On • For research to house lists, make sure
Email List
that lists are double opt-in
Not on
•
Unsubscribe List For general market research, do not
Not Flagged as
buy lists
Spam
• Strongly consider using external panel
Message
Opened
from a reputable provider
Link
Clicked
24
25. The Third Step: Craft Your Invitation Carefully
On • Maintain a suppression list of
Email List
addresses that have opted-out
Not on
Unsubscribe • Provide different categories of
List Not Flagged as
Spam
opt-out (e.g., marketing email vs.
newsletters vs. surveys)
Message
Opened
• Remove suppressed addresses
Link
prior to emailing Clicked
25
26. The Third Step: Craft Your Invitation Carefully
On CAN-SPAM Compliance
Email List
• Truthful subject lines
Not on
Unsubscribe List • Valid “From” e-mail
Not Flagged as
Spam address
• Message
Physical address
Opened
• Valid opt-out link
Link
Clicked
• Avoid SPAM filter trigger words
(“Free, “$”, “act now”)
26
27. The Third Step: Craft Your Invitation Carefully
On • Personal email address
Email List rather than corporate
Not on mail box
Unsubscribe List
Not Flagged as
Spam
Message
Opened
Link
• Compelling subject line
Clicked
o “Help us improve our
products”
o “Share your opinion”
27
28. Clearing the Final Hurdle: Craft Your Invitation Carefully
Less than 8 seconds to make a first impression
On
Tell recipients
Email List what you want them to do as soon
Not on
as possible in the invitation
Unsubscribe List
Not Flagged as
Know your audience and write the invitation with
Spam
this knowledge in mind Message
Opened
For transactional surveys, keep it timely Link
Clicked
For random samples, mentioning the exclusivity
of the invite improves response rates
Let them know how long the survey will be
Consider using HTML e-mails vs. text
28
29. Hurdles in Race to Turn Recipients into Respondents
On A well-crafted invitation is
Email List essential to jumping response
Not on hurdles.
Unsubscribe List
Not Flagged as
Spam
Message
Opened
Link
Clicked
29
32. Clearing the Final Hurdle: Six Points to Cover
Intro/Basic Appeal Your You've been selected for a research
OpinionCounts! study!
Importance of Your participation in this survey Your participation in this survey will
Participation will help improve products and help make improvements to a
services. government program.
Survey Subject We'd like to know your thoughts We'd like to know your thoughts
Matter about gasoline prices. about dairy products.
Flexibility/ We only need to ask you about This survey will only take a few
Time Burden 10 questions. moments of your time.
Incentive Offered Participate and $5 will be Get a free copy of the results after
for Participation credited to the credit card of you participate!
Source: CMOR
your choice
Privacy/Data As member of the Better Your privacy is important to us, your
Confidentiality Business Bureau we take your answers will be combined with
privacy seriously and will others, and will never be linked with
respect your confidentiality. you personally.
32
33. Common Elements of Email Invitations
Incentive amount
Expected length
Contact info
Opt-out/Decline
Subject included
Privacy Policy
Deadline
Panel reminder
0% 20% 40% 60% 80% 100%
Source: Vovici review of 30 ESOMAR 26 documents
33
34. Invitations: To Disclose or Not to Disclose
100% “The invitation text does not
Exclude
90% include…the topic of the
Include
80% survey, as this may affect the
70% panelist’s willingness to
60% participate and thus create
50% bias.” - Norstat
40% “Research shows that people
30% may be more likely to
20% respond when the invitation
10% references a local geography,
0% Normalized. a topic of interest, or an
appealing reward. However,
care must be taken not to
Source: Vovici review of 30 ESOMAR 26 documents bias the responses.” - SSI
35. Effect of Topic & Text of Invitation on Response Rate
• Participate and earn a free gift card from Home Depot
• Participate and win one of 3 cash prizes
• Participate and $5 will be credited to the credit card of your choice
• We’d like to know your thoughts about gasoline prices •This survey is super-short! (we promise)
• Shorter than the typical survey! Takes only 10 minutes… • We only need to ask you about 10 questions
• Your answers will remain completely confidential • We only need to ask you about 25 questions
• We’d like to ask you questions about home improvement
• We’d like to ask you questions about Space Tourism
• Get a free copy of the results after you participate!
• The results of the survey will be published in the New York Times!
• Earn frequent flyer miles on the credit card of your choice
Source: CMOR, “Creating a Respondent Cooperation Playbook”
35
36. Different Constituencies Respond Differently
American College of Cardiology Case Study
Surveys sent between 10 am and 2 pm Tuesday through Thursday
Discovered by analyzing response patterns to past surveys
“Strike at the best time, even if it means waiting.”
Timing of invitations
What works well for consumers may work poorly for businesses
“Is Sunday the new Tuesday?” for B2B surveys
A staggered series of reminders makes optimizing delivery time less important
A/B test
Adapt the split tests used by marketing departments to pit invitations against one
another
Rinse and repeat to find what works best
36
38. Good Reports Require Upfront Planning
Sampling’s Impact on Reporting
Creating questionnaires that facilitate results.
Construct Reports
Easily consumable reports that accurately detail trends and
highlight conclusions.
Present & Share Reports & Data
Efficiently format and share your detailed reports and raw data
38
39. Qualitative/Quantitative Research
• Random samples • Convenience samples
• Closed questions Quantitative Research • Polls with comments
• Limited response options • Online Communities
• Can’t get the story behind • Focus Groups
the story • Provide the Story behind
• Provides hard data that Qualitative Research the story
can be extrapolated to a • Can’t extrapolate to a
larger audience larger audience
39
40. Ramifications for Reporting
Sometimes lots of numbers are really qualitative research,
because you can’t extrapolate from them
Forget sophisticated reporting techniques if research is based on
a convenience sample
But… you don’t need a probability sample to learn that a single
customer is unhappy
40
41. Structure the Questionnaire for Good Data
Preload as much respondent demographic
data as possible
Use conditional logic and piping to keep
respondents from answering questions that
aren’t applicable to them
Use validation to ensure data is provided in
the correct format
41
42. Structure Questions to Fit Analytic Models
Set up parallel questions if you want gap analysis
between importance and satisfaction
Include questions to support the quadrant analyses
and segmentations you will want to report
When designing recurring surveys, develop multi-
question indices to report on
42
44. Use Test Data Before Fielding
Populate the survey with random test data
Generate reports on the mock data to verify your questions will
give you the information that you are seeking
44
45. Understand Your Audience
Who, What, When
Who will be looking at your data?
What are their expectations? Format, delivery,
depth.
How often?
45
46. Building your Report
Use a top-down approach
Start by understanding the major themes, then drill down.
Use filtering and cross-tabulation to find hidden patterns and trends.
Preloaded historical and demographic data will provide essential context.
Do others reach the same conclusions when they work with the data?
Let the results shape the report
Question order should not determine reporting order
Abandon preconceived notions/self-fulfilled ‘realizations’/dramatic over-
generalizations
However, understand, explain and develop conclusions that seem contradictory or
illogical
46
47. Logically Consolidate & Summarize
• Keep it simple and assume nothing.
• Be precise; remove platitudes; use examples and draw logical
parallels
• Include: an executive summary, your methodology, any special
notes. Always cite all sources.
• Call out significant findings. Use the “tell them, then tell them
again” approach.
• Avoid unintentional conclusions. Reconcile easily
misunderstood or contradictory results.
• Provide links to secondary research and articles that:
• Provide perspective
• Reinforce conclusions
47
48. Sharing Your Insight
Build Reports that are Scalable & Repeatable
Researcher’s quandary: People always want more.
Provide a method for other business users to work with reports
Real-time self service & Raw data (filtered or in its entirety)
Choose The Right Format & Distribution Method
Self-service vs. email vs. print
Word, Excel, PowerPoint, PDF, etc.
Consider the Future
The most cost effective research is that which you have already done
Integrate data into future research
48
49. Create Hierarchical Reports
Hierarchical reports eliminate manual processes and offer personalized data per report
user
Case Study:
Corporate Climate Survey, reviewed by 1000 individuals, with different views:
Branch manager and assistant branch manager: branch results
compared to district, region, nation
District manager: each branch in district compared to
district, region, nation
Regional VP: each branch in district, each district in
region, compared to nation
Corporate executives: national results; drilldown to any level
49
50. Five-Part Essay vs. Inverted Pyramid
Introduction Conclusion
Essential
Narration Information
Important
Information
Affirmation
Supplemental
Information
Negation
Data
Tables
Conclusion
50
51. Reports & Presentations in Opposition
Reading a Report Attending a Presentation
Not scheduled in advance Scheduled
No agenda Follows agenda
Private Social
No interaction with author Interaction with presenter
Competing demands on reader No competing demands on attendee
Recommends first Recommendations last
Source: Lawrence Gibson, Eric Marder Associates
51
54. Bar Charts are Easier to Interpret than Pie Charts
Question 1 Question 2 Question 3
A B C D E A B C D E A B C D E
55. Bar Charts Often Inconsistent
% Completely % Completely Likely to
Satisfied Repurchase
30%
Product B Product B 35%
43%
Product D Product D 37%
0% 20% 40% 60% 0% 20% 40% 60%
% Completely Likely to % Completely Likely to
Recommend Expand Purchases
Product B Product B
Product D Product D
0% 10% 20% 30% 10% 12% 14% 16%
57. Common Mistakes with Trend Reports
Showing one year of data
Grouping data by short time periods
Using trend reports at all
Including recent responses in trend
reports
57
60. Use of Faces in Surveys
A face of an interviewer or researcher will increase
the accuracy of reporting
Accuracy is maximized when the gender of the
respondent and the displayed face match
Male panelists are more likely to report accurately
when the introduction is formal, rather than
friendly
Accuracy differed depending on the interaction of
gender of the respondent, presence or absence of a
face, gender of the face and style of introduction
Source: NPD Group
60
61. Continuum of Rich Media Offerings
More like More like
face-to-face paper
interviews surveys
Source: Gina K. Walejko
61
62. Continuum of Rich Media Offerings
Avatars with speech
Faces with recordings
Faces
Surveys routed to human
More like interviewers
face-to- Automating traditional F2F
face probes
interviews
Reintroduces social desirability
bias
62
63. Rich Media Can Increase Completion Rates
Source: Deborah Sleep & Jon Puleston, Quirks – December
2008
63
67. Inference vs. Instruction
42% did not
read the
question
33% did not read
the "traditional"
question
Source: Element54
67
68. Equal Ratings for ‘Engagement’ Metrics
1.0 2.0
Survey
Enjoyment 8.20 8.25
Ease of Use 8.86 8.85
Desire to
Repeat 7.02 6.56
Source: Element54
68
69. Tips for Migrating to Rich Media
Discretion is the better part of
valor
Split ballot test rich media
questions against traditional
questions to identify differences
Visual choices introduce bias,
especially for more abstract
choices
Use rich media in a considered
manner
69
72. Benefits of Multi-Language Surveys
Increase response rates
Improve data quality
Language-specific surveys
Use country as a basis for reports and conditional
branching
Highly efficient means of conducting a single
survey among a global customer base
Allows you to consistently communicate with and
gain feedback from global customers
72 72
74. Study Design
What languages do you need to
translate to?
What internal and external
resources will you use for
translation?
Once the results come in, how
will you translate open-ended
comments?
74 74
75. Remember: Countries are Not Monoliths
Most countries have multiple
languages
There is no Belgian language
There is no Chinese language
Even when two countries share a
common language, you may need to
budget for “localization”
75
7
76. Biggest Mistake in Multi-Language Studies: Unrealistic Schedules
• Prepare a conservative schedule
Study Design
• Time-consuming translation phase after the master
Survey Design questionnaire is complete
• International holiday schedules can extend the time the
Fielding survey is in the field
• Translation of open-ended comments adds weeks to the
Analysis analysis phase
76 76
78. Write Master Questionnaire with Translation in Mind
Using your typical questionnaire would
Increase translation costs
Reduce representativeness of results
Instead, structure for translation
Avoid jargon, slang and technical terms
Rewrite for readability
Use as few open-ended questions as possible
Use language-specific skip patterns
78 78
79. Some Questions Need Alternates, Not Translations
Think localization, not just translation
Country-specific questions are often called-for
No sense translating the choices of the question,
“Which hotel did you choose?”
Choice lists may need to be tailored for major
brands in different markets
Entire sections of the questionnaire may differ
for structure of industry in that country
Graphics and icons may be inappropriate in
some countries or for some cultures
79 79
80. Translation vs. Localization
English
United States
Spanish
French
English
English Spanish Master
Canada
Questionnaire
French
Japanese
Mexico Spanish
Japan Japanese
80
81. Areas of Special Interest
Weights and measures
Currency
Time – 24 hour clock, GMT, Zulu, UTC,
WET
Use images without text to best use in
many cultures
May need to reduce questionnaire
length to take into account overall
budget for translations
81 81
82. Generational Considerations
Use Age Groupings rather than Generational Groupings for
reporting across cultures
Exception: Generational Groupings may be used for the US and Canada
What is Generation X?
Group born between 1964 and 1975 (US)
Younger Westernized generation (India)
Generation Golf (Germany)
Génération Bof (France)
Generation Niks (Netherlands)
82 82
83. Scales Matter
Some scales don’t translate well
They become too wordy, too subjective
Some scales exaggerate cultural issues
Likert scales can be particularly difficult
Safe choices
ConstantSum
MaxDiff
83 83
85. Response Styles by Country: Informed by Culture
Source: Johnson, Kulesa, Cho, Shavitt, 2003; Vovici
85
86. Constant Sum Example
“Please distribute 100
points amongst the Item Points
following 5 items, based Optical zoom of 3X or 30
greater
on their importance to you Resolution: 10 40
when buying digital Megapixels or greater
Available in 5
cameras. An item that has multiplecolors
no importance to you Battery Life 25
should be given no points. Weight 0
100
The total must equal 100
points.”
86
88. Important Safety Tips for Multi-Language Surveys
Translation is the last part of the survey
cycle
Surveys must be finalized before having
them translated to avoid confusion and
re-work
Do not trust machine translations!
88 88
89. A Survey Translation is Not Just a Translation of the Survey
Invitation and reminder emails
Introductions
Button captions
Validation messages
End pages
Thank you messages
Alerts
89 89
90. Choose Your Translator Well
The best translators have excellent
understanding of the originating language, the
target culture, the industry and survey
research
Too often, translation is “insourced” to
employees who do not have the requisite
experience
Survey translations are more demanding than
traditional translations
Use only native speaking professional
translators (preferably certified translators)
90 90
91. Best Practices for Working with Translators
Brief the translators on the intent of the
survey
Create a glossary of English terms and the
target language for consistency
Use native speakers to review the translation
Kathleen Bostick
and provide feedback VP Global Marketing
Kathleen.bostick@lionbridge.com
@Lion_MktgVEEP
Pilot test of translation of survey may be
http://blog.lionbridge.com/
required and will lengthen the process
www.lionbridge.com
A backtranslation may be recommended
Source: Lionbridge
Choose a tool that saves the question and Technologies
translations.
91 91
93. Deadlines and Timing
Timing can affect survey response rates
Be knowledgeable of the holidays and cultural
norms of your survey populations
Avoid survey deadlines on holidays or weekends (in some
cultures Friday/Saturday constitute the weekend, not
Saturday/Sunday)
Be mindful of the International Date Line and individual times
zones within countries
Be careful of time indicators – many cultures use a 24-hour
clock
93 93
94. Respondent Language Selection
• Translated email invitations with links toappropriate translation
• Or self-selection based on all available languages:
94 94
95. Respondent Language Selection
Auto-detect based on primary
language set in browser
Conditional logic, alerts and triggers
based on language-specific surveys
Respondents’ browser settings may
not match their native language
95
9
97. Analysis Considerations
Some countries skew towards “positive” responses (Latin
America, Japan)
Others skew negative (Germany)
Take special care with Product Concept Testing and Brand
Perception Surveys
Always analyze data by country, not just by region
Even if you plan to aggregate by region for reporting purposes
Within some countries, important regional differences are also worth investigating
97 97
99. Optimal Data Storage for Survey Results
French
French
Italian
Italian
German
Base German
Survey
Spanish
Spanish
English Chinese
Chinese
Japanese
Japanese
99
100. Optimal Data Storage for Survey Results
French
French
Italian
Italian
German
Base German Response
Survey Database*
Spanish
Spanish
English Chinese
Chinese
Japanese
Japanese
*Safe
Harbor
Certified
100
101. Optimal Data Storage for Survey Results
French
French
Data
Export
Italian
Italian
German
Base German Response
Survey Database
Spanish
Spanish
English Chinese
Chinese
Japanese
Japanese
101
102. Optimal Data Storage for Survey Results
French
French
Data
Export
Italian
Italian
German
Base German Response
Survey Database
Spanish
Spanish
View
English Chinese Report
Chinese
Japanese
Japanese
102
103. Translation throughout the Survey Research Process
Study Survey
Fielding Analysis
Design Design
103 103
104. What is the Role of the Translator Seat?
User role that provides access to translate specific surveys into
specific languages
Economically allows translators to provide value
Provides security by limiting what the translator can see and do
Option to export surveys as XLIFF file, an XML version of the
survey which translation houses can rapidly use to translate a
single survey into a variety of languages
104
105. Bad Translations: More Than Just Bad Data
Significant schedule slips
Due to re-fielding,
Significant budget overruns new translation
Non-blind surveys: Your brand loses
credibility
105 105
106. Thank You For Your Time
Feel free to email questions:
Jeffrey Henning, PRC
VP of Strategy & Founder
Vovici
jhenning@vovici.com
http://blog.vovici.com
106
It's the desire to better understand the factors which can drive our confidence in the data - which has driven 2 large scale R&D initiatives in 2009.The initial study - "Sexy Questions, Dangerous Results" was presented at the Jan 09 NetGain 3 conference in Toronto. I am extremely grateful to the ResearchNow team, led by John Visser and Rasheeda Qureshi for their partnership on that study. We learned that by moving from a radio button, to a slider or drop down format - we would see significant differences in how often people say they brush their teeth ... how important issues are to Canadians ... and when they last bought products.
what level of inference would we be OK with on the "are you ..." question?
In our Quali/Quant design of 50 interview per interface - respondents rated both experiences similarly across enjoyment, ease of use, and desire to repeat.There are questions which need to be answered - like what will happen over time as survey takers experience the evolution of panels, and overall web experiences.
Generation Bof (Generation Whatever)Generation Niks (Generation Nothing)Baby BoomersUS – between 1946 and 1963Japan – between 1947 and 1949Russia – called Sputnik Generation because they grew up under communism, reachied middle age during the Gorbachev Revolution, and have lived through the transition to a Russian-style market economy.When working in an international environment, you should try and understand the research best practices of the regions you’re working with, when the local agencies usually don’t report generations neither should you probably. And take some history lessons: Understanding a coungry’s past and development phases will help you interpret respondent’s reactions better. Finally, nothing beats talking with locals: speak, whenever possible, to people from the countries you research to understand the nuances of their culture, tradition, and politics.
AttributesA competent translator has the following qualities:familiarity with the subject matter of the text being translated;a very good knowledge of the language, written and spoken, from which he is translating (the source language);an excellent command of the language into which he is translating (the target language);a profound understanding of the etymological and idiomatic correlates between the two languages; anda finely tuned sense of when to metaphrase ("translate literally") and when to paraphrase, so as to assure true rather than spurious equivalents between the source- and target-language texts.[30][edit] MisconceptionsA common misconception is that anyone who can speak a second language will make a good translator. In the translation community, it is generally accepted that the best translations are produced by persons who are translating into their own native languages,[31] as it is rare for someone who has learned a second language to have total fluency in that language. A good translator understands the source language well, has specific experience in the subject matter of the text, and is a good writer in the target language. Moreover, he is not only bilingual but bicultural.
Identify words you don’t want translatedBest to use native speakers in country since usage changesBack translation is especially important in medical or psychometric researchValidation = pilot test of the surveyMay need local Subject Matter Experts to reviewGallup does backtranslations of all their surveys: Gallup does this for all its surveys, having a different person, using a native English speaker to do itMore on backtranslation:
Understand the time differences that may separate various language populations
Skewing toward positive and negative - This is especially critical to plan for in Product Concept Testing and Brand Perception Surveys