Quality in risk reporting on energy issues in German news media

Abstract

In media coverage, the context of the German energy transition, also referred to as Energiewende, casts a spotlight not only on various technology options (e.g. wind or coal power) but also on more abstract topics such as security of supply or electricity prices. Thereby, the public’s assessment of energy-related issues may greatly rely on perceived risks. Focusing on the quality of energy-related risk reporting, this contribution therefore is intended to explore the German print and TV media discourse on energy options or topics. In our sample, one in three articles connects an energy option or topic with an evident or potential unwanted event. Although the media’s effect on actual risk perception involves some controversies, researchers tend to criticize the media as being susceptible to framing and for failing to place unwanted events in perspective, e.g. by not presenting the corresponding likelihood of occurrence, which is necessary to define risk. If this critique holds true, accurate public risk assessment is partly hampered because media coverage implies uncertainty rather than providing all information available. We examined seven indicators of quality reporting derived from literature research: intensity, likelihood, controllability, desired uncertainty, sensationalism, emotional language, and type of unwanted event. Based on German energy media coverage in 2013, we found a relatively high occurrence of intensity and controllability, whereas likelihood and desired uncertainty were reported less often. By aggregating the indicators into a risk-reporting quality index, we did not observe a poor quality of risk reporting on energy issues. In contrast to previous research, the overall quality of energy-related risk reporting can be assessed as at least moderate, implying that the media depicts risks more precisely than assumed. The occurrence of quality indicators thereby significantly depends on the type of unwanted event rather than on the energy option considered.

Keywords:

• risk reporting
• media coverage
• energy technologies
• uncertainty

Notes

1. Own translation from German.

2. Framing means selecting aspects by neglecting others, presenting a specific perspective of an issue (Entman 1993).

3. Heidmann and Milde (2013) identified a strong relationship between the occurrence of uncertainty and risks in the media coverage of nanotechnology.

4. A further problem is that scientists’ numerical assessments often are flawed and imprecise (Friedman et al. 1996), making responsible journalism with a high quality of risk reporting more difficult.

5. Guenther, Froehlich, and Ruhrmann (2015) identified that journalists’ decisions on how to cover (un)certainty in their articles are influenced by their attitudes, social norms, and perceived behavioral control.

6. Energiewende* OR Energiepolitik* OR Kraftwerkspark OR Energiesektor OR Energieträger OR Strombedarf.

7. However, any type of pictures was considered within the evaluation of, e.g. sensationalism or personalization, although they are not treated as a separate unit of analysis.

8. ‘Specific’ refers to an individual energy option, e.g. coal rather than any consideration of fossil or renewable energies in general.

9. Up to three unwanted events could be coded for each energy presentation. For energy options, 125 unwanted events were recorded as a first mention, only nine times a second unwanted event was coded, a third unwanted event never occurred.

10. Because four cells have expected frequencies below 5, Fisher’s Exact Test would have been a better solution. However, due to the complexity of the contingency table, its computation was not possible within the statistics software applied (SPSS).

11. Perhaps conservative newspapers tend to frame renewable energies negatively, and therefore use wind energy as an example due to its controversial nature. Thus, they stress unwanted event-related wind energy reporting, which causes a higher share of wind energy and a higher share of unwanted event-related energy coverage.

12. Seven cells have expected frequencies below 5, see endnote 10 for further explanations.

13. For each presentation, up to three types of unwanted events could be coded, which caused the slightly higher number of presentations in the following charts.

14. Eighty-four cells have expected frequencies below 5, see endnote 10 for further explanations.

15. However, note that an explicit mentioning of uncertainty is not essential in case of an unwanted event already having occurred. In future studies, the degree of the occurrence of uncertainty that would have been necessary for high precision in unwanted event-related energy reporting should additionally be coded.

16. We only considered energy options with at least eight presentations of unwanted events and types of unwanted events with at least 10 unwanted event-related presentations.

17. Three cells have expected frequencies below 5, see endnote 10 for further explanations.

18. Three cells have expected frequencies below 5, see endnote 10 for further explanations.

19. Two cells have expected frequencies below 5, see endnote 10 for further explanations.

20. This was because each presentation can only be assigned to one index score and we did not aim at calculating mean values in cases of more than one unwanted event-coding per presentation.

21. The index ranges from 0 to 7 index points; a midpoint of 3.5 and was chosen as test value for the t-Test.

22. We only considered those energy options having at least five presentations of unwanted events.

23. (Parametric) ANOVA provides similar results: F = 0.870, p = 0.530. Due to homogeneous variances, we also can rely on ANOVA results, although our risk-reporting quality index shows no normally distributed data.

24. (Parametric) ANOVA provides nearly similar results: F = 0.974, p = 0.447.