Quarterly webinar with our CEO - obtain valuable input on survey sampling and data quality

06.09.2021

Survalyzer’s upcoming quarterly review & outlook webinars presented by CEO and Managing Partner Christian Hyka give you valuable insights and the opportunity to stay up to date with the constantly evolving software Survalyzer Next Generation.

Please click on the link below to watch the second Webinar presented beginning of September, 2021:

Webinar – Quarterly Webinar with CEO September 2021

Not interested in watching the webinar? Read the summary below and have a look at the product roadmap:

Further development of Survalyzer Next Generation

We used the Summer brake to develop further our Software Survalyzer Next Generation.
In the 3rd quarter we focused on 2 topics:

  • Valid survey data quality
  • Secured Dynata sampling process

(01:30) Valid survey data quality

Data quality is a big topic for everyone who does online surveys. Survalyzer has now an Artificial Intelligence (AI) based data quality assurance algorithm to make sure your respondents fill out your survey in a valid way.

What quality is covered by our data quality algorithm?

We focus on the seriousness of the interviews, means that every online survey needs certain kind of attention and honesty from the respondent. We can measure this type of quality and identified 4 different types of unserious answer behaviour:

  • Trap questions (e.g. ask: “Are you happy today?” Respondent answers “Yes”, ask same question later again and then get a “No”)
  • Speeding
  • Bad open-end
  • Straight-lining

 

Out of these 4 types of answer behaviour we developed 4 quality scores. Each quality score ranges from 0 (no quality complain) to 100% (bad quality):

  • Trap question score (either caught in trap or not)
  • Survey completion time (current completion time in comparison to median completion time from other respondents)
  • Open questions answers (here we use the Microsoft Azure cognitive services, which is an AI that reads the text and can classify if the input is valid text)
  • Response behaviour (check if respondents gave always the same answer option)

With the following 3 rules we can identify 95% of bad responses:

  • Double trap (participant falls for at least 2 traps)
  • 2 out of 4 (if 2 of the quality scores are 100%)
  • 240% (none of the quality scores is 100%, but all together are equal or above 240%)

 

Our investigations on the raw data showed, that the bad responses are significant and build the majority to identify bad quality. Bad responses are the best indicator for a bad quality score.

What is the added-value when you use this automated quality assurance algorithm?
  1. You don’t have to manually control your data, so in the end you will save time and costs.
  2. You gain precision because of the automatic quality screenout, so you don’t need to oversample (order more than you actually need from a field agency).

(09:30) Secured Dynata sampling process

We are going to roll out this quality assurance algorithm over all type of surveys, but we start now with this algorithm in our new Dynata API-integration.

Survalyzer added a new interface to Dynata, that means we have developed an integration to the panel provider. Dynata is a field research agency and provides you a target group if you don’t have your own list of respondents.

How do you launch a study with a field research agency in Survalyzer?

When you want to launch a study with a field research agency, you need to have different sections in the survey, one is the screening section.

 

The screening sections contain age questions, gender questions, different trap questions and the scripts (value assignment). The quality score itself is reflected by these scripts, this means the output is used in the screener section of Dynata. Don’t worry, you do not have to build those kind of complex scripts, we will provide you all the templates.

 

The survey can be then launched in the Distribute tab of your survey. There we have a new tile for the Dynata sample.

 

By clicking on the tile, you open a new wizard to buy your target group. We offer two types of wizards:

  • The wizard shown in this summary is the one with all steps for our Professional users.
  • For our Essential users we offer a a simplified one to easily define and buy their target group.

In a first step you need to:

  • enter the name of your sample
  • define the number of interviews you want to order
  • specify in which country
  • specify the topic of the survey (not really relevant for you, but for the field agency)

 

In the second step you can select the target group for your survey. You have a very broad range of different attributes.

 

In the next step you can segment the sample, which means you can split the members equally by creating quota groups.

If you for example want to make sure that you have an equal distribution between different age categories or between genders, you can create appropriate quota groups and give Dynata the command to segment according to those quota groups.

 

In the screenout section you have to define, which participants you do not want to have in your survey. The speeders and the straight-liner are automatically screened out as well as the bad open ends. The trap questions need to be configured.

 

In the fifth step you have additionally to the cloud quotas the possibility to add survey quotas on the survey questions. And in the sixth step you have to specify the field phase and how long the interview is (the precalculation is done by Survalyzer).

The incidence rate has also to be estimated. If you are not really familiar with this topic then keep it at 100% and make sure you do not have too narrow screenout conditions. Survalyzer can also advise you on the incidence rate estimation.

 

Then we can make a feasibility check. That means:

  1. We send all the configuration information to the field agency (in this case Dynata)
  2.  They look if they can deliver this sample
  3. What is the price

 

In the last step you get a final overview of all the configurations. If you are fine you can accept the terms and conditions and make a purchase. The invoicing is then done over Survalyzer.

What’s coming next (coming quarter)

(24:12) Roadmap and outlook of the coming quarter

You can find the roadmap with an overview of what new features have been delivered and what we’re currently working on in our support center.

In the next 2 weeks we are going to release the Dynata integration shown above.

The following 4 topics are currently under development:
  • Report Builder Pro: Available report builder in Analyze area. This is a big topic and we will tell you more about this in our next webinar.
  • NPS question type: This is actually already now possible in a data collection point of view, but we are going to extend the analytics part to show the NPS KPI directly in the dashboard.
  • Question library: In the library version 1.0 you will have the opportunity to save the questions you have created and customized before and to reuse it in another questionnaire (question template).
  • Change question type: This function makes it possible to change an already created single choice question to a multiple choice question for instance, so you do not have to delete and recreate the question. The choice text and question text will automatically be transferred to the new question.

Other topics

(26:30) Our new project manager

We are delighted to announce our newest employee Milan Aleksendric. He has multi-year experience in IT project management and will support Christian Hyka in all integration projects.

Next webinar with our CEO

Save the date for the next webinar with Christian Hyka on December 02, 2021.

We look forward having you at our upcoming webinar!