"there used to be more Hitler..." (reference)
an analysis of "Der Spiegel" covers over the course of time
This is a rather quick translation of my original article which I published in German on September 1st. Please excuse the German headings on the graphs!
Since 1947 the SPIEGEL, an "assault gun of democracy" or a "former news magazine" (according to readers opinions) has created more than 3500 covers. The shown collection of covers represent a quiet exciting historical archive: Covers reflect important social issues and political discourses of their time. I looked for trends inside this archive and classified each cover into the categories "man", "woman", "other", "Hitler" and "sex-sells" over a longer time period. Sometimes this categorization is somewhat blurry. Hence the data presented here is certainly correct in terms of general trends but might be controversial in individual cases. This is an after-work project, please excuse any mistakes I may have made. To ensure verifiability of the evaluations I hereby publish the raw data and am open to feedback.
an unbalanced gender ratio
Until the 1960s the covers usually consisted of portraits of (mostly male) people. At first this may seem little surprising, but it is interesting to see that this pattern was even more obvious in the years 1955-1962. Although the proportion of all-male covers has declined significantly in the following decades, the proportion of women was not considerably increased. This comes as a consequence of a trend to use cover images not showing people which started in the 1960s. I should point out the way I classified the covers though: I put photos and more or less photo-realistically drawn living people into the categories "man and woman" while I categorized rather abstractedly drawn people and small children as "other".
"sex-sells" & Hitler - clickbaits are not a new phenomenon
In recent months there was an increasing debate on the concept of "ClickBaits" being used by media outlets. This term refers to article headlines that appeal to specific stimuli and animate the user to click or share a link. "You will not believe what happened next!" In fact, the utilization of certain key images and concepts is not a new phenomenon at all. Thus the formula "sex-sells" is probably as old as the media industry itself and of course I added this as a category. Additionally I classified the covers showing Hitler as another category thinking they might be an example of provoking negative reactions. Of the 3512 covers 48 covers could be filed in the Hitler category and 152 cover were classified as "sex-sells". It is interesting to see the frequency of these supposedly "BuyBait covers" over the years. Three aspects can to be mentioned:
- Between 1954 and 1963 there were no "sex-sells" - or Hitler covers
- "Sex Sells" has had two peaks: One in the 1970s and another in the mid / late 1990s
- Guido Knopp's favorite theme of German history (Hitler) seems to correspond to the second "sex-sells" peak
The first peak of "sex-sells" covers in the late 1970s is probably due to the arrival of the sexual revolution in the minds of the makers of the journal at that time. But how can the very similar rising and then falling graph of both curves in the mid / late 90's be explained? A very plausible theory would be that in 1993 the Spiegel got a new competitor - the Focus magazine. Maybe this influenced the tabloidisation of the covers. This situation however, significantly declined in recent years. An analysis of the Focus covers would of course also be exciting. Perhaps another time.
"sex-sells" - even back pains
I tagged the "sex-sells" covers regardless of the gender of the person depicted. And although the evaluation provides no real surprises, I created a graph showing the gender distribution on the covers for this category. More than 50% of the cover photos solely show women, more than a quarter of the covers show male-female couples. Technically the proportion of women is even a little higher, because drawn people were classified in the gender category "other".
Leafing through the covers I realized again how often medical issues are actually illustrated by women. Unfortunately, I have made no explicit evaluation on gender vs. topics, but the probability that the topic of "back pain" is illustrated with a naked woman, seems to be at 95%.
so does sex sell?
Has the use of the "sex-sells" covers affected the economic success of an issue? I was wondering whether those schemes really work. Fortunately, the IVW offers subscription-adjusted sales (retail sales) as an csv download. Due to the declining sales of the past years I calculated the arithmetic mean values for the "sex-sells" covers and for the rest of each year individually. The result is somewhat ambiguous: During the years 2005, 2006, 2007 and 2009 the "sex-sells" covers seem to have resulted in an increase in sales, whereas in the years of 2008 and 2009 they resulted in a little decrease.
looking for the silly season
Once you dig a little deeper into the dataset you find some interesting facts: I created the following chart showing the distribution of sum of "sex-sells" covers for each edition number per year. With a little good will one might get the impression that the silly season from late May to August is more frequently occupied with the "sex-sells" category than the rest of the year. The amount of data on Hitler covers was not sufficient to come up with a meaningful evaluation.
Some last questions:
Since this is an after-work project of course a number of questions still remained unanswered. I leave these questions up for discussion including the raw data and would like to encourage everyone to continue my work on this topic:
- How does the gender ratio compare to the covers of other magazines?
- Reclassification of all cover by topic. How are the subjects of politics, sports, medicine head lines?
- What other forms of BuyBaits can be identified? Certain words, colours, typography... ?
- How often are certain topics repeated and with which attitude?
* Disclaimer: I work part-time as a data journalist for Spiegel Online.