Case Study - Comparison of classic and Redem cleansing in the case of Magenta Smart-City

In this paper, we shed light on the effects of the Redem adjustment on the interpretation of the results of the Magenta Smart City study.

The study was conducted by TripleM – Matzka Markt- und Meinungsforschung on behalf of Magenta Business carried out and cleaned up by Redem.

In a comparison of conventional and Redem cleaning, we impressively show the effects of cleaning by Redem on the interpretation of the data. This is especially important for recommendations for action in the case of more complex questions.

Survey design | Description of the survey

For the cleaning of the raw data of the Magenta Smart City study, the conventional cleaning of speeders, straightliners and also the control of open data was first carried out manually by TripleM.

Subsequently, the conventionally cleaned data was analysed and weighted using the "redem prediction score", which measures superficial response behaviour through projective control questions.
This method identifies respondents whose answers were given carelessly and without cognitive effort.
Especially with more complex and longer questions, this effect can lead to negative distortions and can hardly be corrected by conventional adjustment.

The following excerpts from the Smart City study show the effect of the Redem clean-up compared to the conventional clean-up.

Weighting - How social demographics are taken into account

The social demography is hardly changed by the Redem adjustment.

The data was socio-demographically weighted and combined with the Redem weighting factor. With socio-demographic data, a superficial answer is not to be expected.

Awareness of the term "Smart City"

In terms of term awareness, the redem adjustment brings a slight increase.

Here, too, no great change was to be expected from the Redem adjustment, since it is a relatively simple question.

Importance of the municipality/city becoming a smart city?

The importance of Smart City for one's own municipality would be significantly underestimated WITHOUT the Redem purge.

Without the cleaning by Redem, there would be a 'differentiated interpretation of the data. The importance increases by over 10 percentage points, to 75,9 percent.
This question is an important element of the entire study and essential for a valid result.

Due to the more complex content of the question, the effect of the redem adjustment is very clear.

Assessment of Smart City - agreement with statements

The adjustment by Redem shows partly significant differences in the assessment of Smart City.

This example also shows that superficial response behaviour is very likely, especially with questions where respondents have to make an increased cognitive effort.

Increasing the quality of the sample

Redem adjustment increases the quality score (R-score) of the whole sample by 14 Scale points on 81 Punkte.

The Redem score (R-score) measures the data quality, with a value between 0 and 100, before and after cleansing by Redem.

Conclusion

In the Magenta Smart City study, superficial response behaviour occurred primarily with more complex questions, with increased cognitive effort.

This case thus underlines the importance of red-emphasis and shows very well which questions are likely to be biased by superficial answers and which questions are less susceptible to it.

Click here for the press release of the study:
https://newsroom.magenta.at/2021/08/05/magenta-smart-city-studie/