Vertebrate species knowledge: an important skill is threatened by extinction

ABSTRACT

Vertebrate species knowledge, a key factor for conserving biodiversity, is a compulsory part of school curricula. This study evaluated the vertebrate species knowledge of sixth-grade German grammar school pupils (aged 11–13) to investigate the influence of socio-demographic factors (gender, size of hometown, mother tongue), personal expectations (favourite school subject, performance in biology, expected test results) and possibilities to do nature observations (favourite playing area, number of excursions to nature) on identification skills. In addition, the study examined whether schoolchildren's taxonomic knowledge changed over the years. In a species identification test, participants (N = 984) scored an average of 14.18 ± 3.82 out of 25 points. The results showed that their species knowledge was significantly influenced by their personal expectations on test performance, the school, favourite playing area, mother tongue and gender. Hometown size, the number of excursions to nature, favourite school subject and performance in biology did not significantly affect children's species knowledge. Mammals had the highest identification rates, whereas birds and reptiles were rather unknown. These results were compared with a similar dataset in 2006 indicating a 15% loss in pupils’ species knowledge within the last decade. A change in the curriculum as a reason for this decline in taxonomic knowledge is discussed.

KEYWORDS:

• Species knowledge
• biodiversity
• conservation
• biodiversity awareness
• science communication

Introduction

Why do children need species knowledge?

There is a long history of warnings that humans are facing a mass extinction caused by themselves (United Nations, 1992). The loss of biodiversity recently riveted the attention not only of some ecologists, but a broad part of society (Díaz et al., 2019; Hallmann et al., 2017; Steffen et al., 2015). Even the World Economic Forum (2018) listed the ‘loss of biodiversity’ among the 10 most dangerous threats for the welfare of societies. Taxonomic knowledge is a fundamental skill when it comes to understanding ecological connections and the conservation of biodiversity (Bilton, 2014; Leather & Quicke, 2010; Palmberg et al., 2018; Randler, 2008). Therefore, it is vital to have taxonomic experts, but most taxonomic professionals are rather old (Frobel & Schlumprecht, 2016; Kuhlmann, 2015). The possible extinction of experience might cause a shifting baseline problem relevant to nature conservation activities (Miller, 2005).

It is also vital to teach children species knowledge because of the positive correlation between the understanding of nature and the disposition to preserve natural resources. Studies of the past showed that people with high species knowledge are more willing to save the environment than others (Hosaka et al., 2017; Prokop & Fančovičová, 2013; Scott et al., 2012). From the society’s point of view, species knowledge is therefore essential for developing conservation strategies by taxonomic experts as well as for the acceptance of these strategies in society.

Many curricula demand the teaching of species identification skills. They are present in the sustainable development goals of the United Nations (Rieckmann et al., 2017), as well as in compulsory curricula of schools (e.g. Palmberg et al., 2019; Staatsinstitut für Schulqualität und Bildungsforschung, 2017). Teaching biodiversity at school is not only an obligatory necessity but offers many possibilities for hands-on activities (Fančovičová & Prokop, 2017) or outdoor experiences (Lindemann-Matthies, 2006) to motivate pupils and enrich biology lessons.

A closer connection to nature offers further educational benefits, e.g. increasing student’s attention to nature (Li & Sullivan, 2016). Furthermore, acquisition of species knowledge fosters children’s ability to detect small morphological differences, which is vital for the development of their observational skills from early childhood (Kohlhauf et al., 2011) to formal educational processes in schools (Tomkins & Tunnicliffe, 2001) and to professional qualification procedures such as medical diagnostics (Jasani & Saks, 2013).

There are also other strong motives to teach species knowledge for the personal wellbeing, as sound ecosystems have a positive effect on the physical health of humans (Lindley et al., 2019). Since Louv (2008) coined the term ‘nature deficit disorder’, many surveys have shown positive correlations between biodiversity and mental health (Cox & Gaston, 2015; Keniger et al., 2013; Marselle et al., 2019).

Determinants influencing species knowledge

Socio-demographic factors

Gender: There are contradictory research findings on whether boys or girls know more animal species. In some studies, boys achieved higher scores than girls, but these studies are either quite old (Eschenhagen, 1982) or conducted in far away countries (Nyhus et al., 2003) and thus the socialisation process is not comparable to our test group. Other studies contained non-native animals (Huxham et al., 2006), so the outcome may be biased. Peterson et al. (2017) found better species knowledge of boys, which was explained by their higher interest in hunting activities.

On the other hand, there are surveys reporting better species identification skills for females. Birds in particular seem to be more popular among girls (Gerl et al., 2018; Zahner et al., 2007), but these researchers explained the findings with the girls’ higher motivation to achieve good results in the test, rather than with their knowledge differences. Yet other surveys could not find any gender-related differences in identification skills (Hooykaas et al., 2019) or general knowledge of animals (Prokop et al., 2008).

Size of hometown: Although several studies reported that the children’s ability to identify plant species is better in rural areas than in big towns (Lückmann & Menzel, 2014; Remmele & Lindemann-Matthies, 2018), it might be different for animal identification skills. Older surveys (Eschenhagen, 1982; Zahner et al., 2007) reported slightly better results for pupils inhabiting rural areas, but recent studies on animal knowledge do not find any differences between rural or urban children (Hooykaas et al., 2019; Remmele & Lindemann-Matthies, 2018) or even better results for pupils in big cities (Gerl et al., 2018).

Children, especially pre-school or primary schoolchildren, need a mediator to communicate species knowledge (Chand & Shukla, 2003; Remmele & Lindemann-Matthies, 2018; Tarłowski, 2019). This person is often part of their family (Remmele & Lindemann-Matthies, 2018; Tunnicliffe & Reiss, 1999). The older the children are, the more important their learning processes in formal education (e.g. biology lessons) should become. As the quality of the mediation process in different classes might vary, the school children attend should influence their species knowledge.

Mother tongue: Several studies showed that immigrant children with another mother tongue are likely to score weaker results in tests on scientific skills (Wendt et al., 2016) due to their lack of vocabulary (Kempert et al., 2016). As many species names are not related to morphological characteristics (e.g. blackbird) but the name of scientists (e.g. Temminck’s stint), they are difficult to memorize (Randler & Metz, 2005) and should therefore be even more difficult to learn for pupils with a different mother tongue.

Animal observations

Favourite playing area: Species knowledge is positively influenced by animal-related activities such as keeping a pet (Prokop & Tunnicliffe, 2010; Randler, 2010). In addition, direct observations of species are an effective strategy to foster pupils’ identification skills (Gerl et al., 2018; Hooykaas et al., 2019; Palmberg et al., 2018; Palmberg et al., 2019; Randler et al., 2015; Remmele & Lindemann-Matthies, 2018). Children who prefer to play in nature have a higher chance to encounter animals and should outperform those with less possibilities to observe species.

Number of excursions into nature: As nature experiences influence children’s attitudes towards ecosystems and in consequence their ecological knowledge (Torkar & Krašovec, 2019), observations of wild animals should also improve their species knowledge. But the chance to see an animal by accident during leisure time activities might not increase children’s ability to name animals as much as during nature excursions in biology lessons. Therefore, pupils who participated in more excursions should have better identification skills.

Favourite school subject and personal expectations

Studies reported that pupils achieve higher test results if the test items are similar to the content of their favourite school subject (Kubiatko et al., 2012; Raza & Shah, 2011). Assuming that these results could be generalised, researchers can argue that pupils who like biology as a school subject should recognise more animals. Furthermore, high personal expectations to perform well in tasks were reported as an explanation for higher test results (Eccles & Wigfield, 2002; Schunk et al., 2014). Pupils who think that they do better than their classmates in the subject biology or perform well in the animal identification test itself should therefore achieve higher scores.

Change of species knowledge over the years

Atran et al. (2004) claimed a deterioration of knowledge and understanding of natural surroundings (including species identification) within several technologically oriented societies, whereas indigenous groups with a closer contact to nature could identify plants and animals better on the species level. These findings are in line with the study that children from western societies know more Pokémon creatures than real animals (Balmford, 2002).

However, empirical data supporting a decline of species knowledge over time is quite rare because there are only a few older studies investigating children’s taxonomic knowledge that could be used as a baseline to examine any changes through the years. First hints came from Randler (2006), who repeated a survey on species knowledge of vertebrates among German pupils done by Eschenhagen (1982). But these two studies were not comparable asEschenhagen used stuffed animals, while Randler (2006) displayed pictures for testing pupils’ species knowledge.

Gerl et al. (2018) repeated a survey of Zahner et al. (2007) with pupils in higher secondary education. The participants in 2017, who had to identify common birds, scored an average of 20% less in the same test than did those back in 2007. In this present study, we want to verify if those findings were just a singular coincidence attributed to birds or a common trend in schoolchildren’s declining vertebrate species knowledge in general over the years.

Aim of the study and research questions

This study evaluated how many native vertebrate species grade six pupils from grammar schools can identify. Furthermore, we wanted to find out if certain sociodemographic factors (gender, size of hometown, mother tongue), animal observations, personal expectations and interests influence the children’s ability to recognise animals. In addition, we tried to answer the question whether species knowledge has declined over the years by comparing the results of today’s children with similar data collected by Randler (2006) more than a decade ago. All examined questions and tested hypotheses are summarised in Table 1.

Table 1. Questions and hypotheses of this study.

Which determinants influence species knowledge?
Determinant		Hypothesis
Socio-demographic factors	Gender	Girls know more species than boys.
	Size of hometown	Children living in rural areas know more species than those inhabiting big towns.
	Mother tongue	Native speakers know more species than those with another first language.
Animal observations	Favourite playing area	Pupils, who prefer playing in nature know more species than those playing indoor.
	Number of excursions into nature per school year	The more excursions into nature children make, the more species they know.
Expectations and interests	Favourite school subject	Pupils whose favourite school subject is biology know more species than others.
	Performance in biology compared to classmates	The higher the performance of the tested pupils in biology is, the more species they will know.
	Expected test results	The higher the expectations of the tested pupils are, the more species they will know.
Does species knowledge decline over the years?
Factor		Hypothesis
	Change of knowledge over the years	Pupils in 2018 know less species than those of a comparable group in 2006.

Materials and methods

Sample

In 2018, a paper and pencil test was conducted with 984 (n_male= 476, n_female= 497) sixth-grade pupils in grammar school. The 40 participating classes from 20 different schools were chosen by personal contact. More than half of the participants (52.7%) lived in small towns with 5000–50,000 inhabitants, about one third (31.9%) came from villages with less than 5000 inhabitants and 11.4% dwelt in big cities with more than 50,000 residents. This classification of different settlement sizes follows the one made by Gerl et al. (2018). Most of the participants (85.0%) spoke German as their first language.

These results were compared with those of 154 (n_male= 65, n_female= 89) sixth-grade pupils in a German grammar school back in 2006 (Randler, 2006).

Measures

The pupils had to identify 25 native vertebrate species by viewing coloured pictures of animals in an open-ended item format. The pictures were checked prior to the test by three taxonomic experts to ensure all necessary criteria for identification are visible. The distribution of species in the test (6 mammals, 10 birds, 3 reptiles, 3 amphibians and 3 fish species) reflects the distribution of native vertebrates among the vertebrate classes (Haupt, 2009) in Germany. The animals were selected because they either appeared in previous tests on species knowledge or are attractive species representative for a special ecosystem. The pupils had 45 minutes to complete the test in the presence of their teacher.

According to previous studies (Eschenhagen, 1982; Gerl et al., 2018; Randler, 2006; Zahner et al., 2007), the participants could score 1 point for naming one species (i.e. a maximum of 25 points) and 0.5 point if they could not report the exact name of the species but a higher taxonomic group. The resulting scale species knowledge had a good reliability, Cronbach’s α = .80 (Cortina, 1993). For comparison of this study’s results with those of Randler’s (2006) study, only 20 of the 25 species could be used, as the Grey partridge (Perdix perdix), Eurasian Eagle-Owl (Bubo bubo), Eurasian Lynx (Lynx lynx), Eurasian Beaver (Castor fiber) and the European Eel (Anguilla anguilla) were not part of Randler's survey. The reduced scale of species knowledge 20 had an acceptable reliability, Cronbach’s α = .77.

To identify determinants influencing pupils’ species knowledge, the questionnaire contained closed format items asking for information about gender, size of hometown, mother tongue, favourite school subject or playing area. Furthermore, we asked for the number of excursions into nature pupils took part during a school year. In addition, the pupils had to rank their interest in the different vertebrate classes to compare this ranking with the identification rates of the taxonomic groups.

To test if the expectancy-value theory (Eccles & Wigfield, 2002) could be used to predict species identification skills, pupils were asked to estimate their performance in the school subject biology compared to their classmates, their expected test results and their interest in different vertebrate classes.

A translated version of the original test instrument with all questions analysed in this study is available for download (Appendix).

The identification rate is defined as the proportion of participants’ correct answers andcharacterises their popularity of the species. These rates were categorised in four quartiles shown in Table 2. In a similar procedure, we calculated the identification rates of vertebrate classes by dividing the average score within a systematic group by the number of species belonging to each class in the test.

Table 2. Popularity of the species according to the identification rate.

Identification rate	Popularity category
100–75%	Well known
74–50%	Rather well known
49–25%	Rarely known
24–0%	Unknown

Data analysis

The average score M on species knowledge was used to describe pupils’ ability to identify animals and to compare this ability between different groups within the participants. All tested variables did not show strong Pearson correlations (r > .35). So a general linear model (GLM Type III SS) with the pupils’ school as random factor and all other explanatory variables (number of excursions, gender, mother tongue, size of hometown, expected test result, preferred leisure time activities and the favourite school subject) as fixed factors, was used to test the influence of these variables on the average species knowledge M. Non-significant variables (p > .05) were eliminated for the final model. The statistical analysis was carried out with IBM SPSS Statistics for Windows version 25. The effect size partial eta-squared (η²) was used to quantify the effect of a determinant on the average species knowledge.

For detecting possible changes in identification skills between the children in 2006 and 2018, Randler’s original data in 2006 underwent the identical statistical procedure as for the data of 2018. Although the scores of species knowledge do not follow a Gaussian distribution using Shapiro–Wilk-Test, D(984) = .995, p < .005, we used parametric tests (t-test) to compare the differences in average scores, because the number of items building the scale and the number of participants were rather high and skewness test results are usually significant with a sample size higher than 500 (Field, 2013). To measure the effect size of the different average test results in 2006 and 2018, we used Cohen’s d.

To find out if the identification rate of a single species had changed over the years, we calculated the identification rate of each species in both test groups separately and compared the two means with a t-test.

Results

Determinants influencing species knowledge

The pupils in 2018 could identify an average of 14.18 ± 3.82 of species of the given 25 species in the test. Table 3 displays an overview of the variables correlated with the participants’ average species knowledge.

Table 3. Factors influencing the average species knowledge M with standard error (SE), the ANOVA-Values (F), significance (p) according to a univariate GLM sorted by the effect size partial η².

Factor		n	M ± SE	F	p	η²
Expected test result	<5	47	8.6 ± 0.65	48.713	<.001	.200
	5–9	154	9.8 ± 0.26
	10–14	285	11.3 ± 0.19
	15–20	336	13.4 ± 3.32
	>20	133	15.33 ± 0.40
School (Figure 1)				4.974	<.001	.110
Favourite playing area	Indoor	147	12.81 ± 0.32	10.46	<.001	.030
	Outdoor	537	14.34 ± 0.16
	Nature	149	15.10 ± 0.29
Mother tongue	German	836	14.51 ± 0.10	19.463	<.001	.020
	Other	148	12.28 ± 0.33
Gender	Male	476	13.97 ± 0.19	8.208	.004	.010
	Female	497	14.41 ± 0.16
Size of hometown	<5000	296	14.13 ± 0.22	2.469	.085	.007
	5000–50,000	519	14.01 ± 0.16
	>50,000	112	15.36 ± 0.38
Performance in biology compared to classmates	Better	162	14.81 ± 0.30	0.861	.423	.003
	Equal	609	14.27 ± 0.15
	Worse	185	13.02 ± 0.28
Number of excursions per school year	0	548	13.76 ± 0.16	0.490	.689	.002
	1	233	14.27 ± 0.26
	2	122	14.91 ± 0.32
	>2	60	16.29 ± 0.49
Favourite school subject	Biology	315	14.19 ± 0.23	0.617	.433	.001
	Other	669	14.17 ± 0.14

Socio-demographic factors.

There were significant gender differences, but the effect size η² was small (Cohen, 1988). The size of the pupils’ hometown showed no correlation with their species knowledge, but the pupils’ school (and therefore their teacher) correlated significantly (Figure 1) with a medium effect size.

Figure 1. Average species knowledge M of pupils in different schools with error bars showing the standard error (SE).

Native speakers achieved higher scores than did those with adifferent mother tongue. The difference of about 1.3 points was significant with a small effect size.

Animal observations

Another significant but small effect on species knowledge is connected to the place where pupils like to spend their leisure time. Those who prefer to play in nature (e.g. forests) scored about 1.5 points more compared to those who like to play outdoors (e.g. sports) and even 2.3 points more compared to those who favour indoor activities (Table 3).

Although the number of nature excursions was not significantly associated with species knowledge according to the general linear model (GLM), we found that the group of pupils who attended more than two of these excursions per year (n = 60) scored an average of 16.3 points, which is the highest result in the group (Table 3).

Favourite school subject and personal expectations Although pupils. who thought that they could do better in biology compared to their classmates. scored higher in our test, there was no significant correlation between species knowledge and their expected performance in biology or between species knowledge and their favourite school subject. The expected test results were found to have a large effect on their test scores, but many pupils underestimated their scores. Figure 2 summarizes the determinants’ correlation to species knowledge.

Figure 2. Summary of factors influencing species knowledge with their effect sizes. Non-significant factors are marked with dotted lines and smaller font size.

Identification rates

As displayed in Figure 3, the group of eight well-known species (indentification rate higher than 75%) consists of mainly of mammals (six species), whereas all species identified by less than 25% were birds.

Figure 3. Proportions of correct identifications on species level tested in 2018. Species that were not tested in 2006 are marked with an ‘*’.

Hence, the results showed that the participants’ best-known group of vertebrates were the mammals followed by fish and amphibians, whereas birds and reptiles were rather unknown. This result did not relate to the interest in the different vertebrate classes stated by pupils (Figure 4).

Figure 4. Proportions of correct identifications for the different vertebrate classes with error bars showing the standard error depending on the pupils’ interest in the vertebrate class.

Table 4 shows differences of 5% or morebetween the rate of identification on species respectively a higher taxonomic level in 2006. These eight animal species were the same in 2018 except for the Common Buzzard. The remaining twelve animal species could either be identified on species level or not at all.

Table 4. Species showing a difference of 5% or greater between the identification rates on species level and on a higher taxonomic level in 2018 and 2006.

	2018			2006
Species / Higher taxonomic level	Diff. (%)	Higher taxonomic level (%)	Species level (%)	Diff. (%)	Higher taxonomic level (%)	Species level (%)
Mallard/Duck	41	57	16	27	68	41
Common Frog/Frog	30	45	15	27	32	4
Common Toad/Toad	28	47	19	25	51	25
Gr. Spotted Woodpecker/Woodpecker	27	64	37	12	52	40
Sand Lizard/Lizard	19	30	12	27	34	7
Great Tit/Tit	7	33	26	12	48	36
Fire Salamander/salamander	6	83	77	13	73	60
Common Buzzard/Buzzard	3	14	12	8	23	15

Change of species knowledge since 2006

Between 2006 and 2018, we observed a significant difference in species knowledge within the test groups. The participants in 2018 (M = 10.86 ± 3.26) had a lower average score which was 3.0 less than that of their peers in 2006 (M = 13.83 ± 3.26). This difference of about 15% over the last decade is marked by a large effect size (p < .001, Cohen’s d = 0.91): for boys – 2.4 less (p < .001, d = 0.8), for girls 3.3 less (p < .001, d = 1.1) (Figure 5).

Figure 5. Average species knowledge (M) of boys and girls in 2006 and 2018 with error bars showing the standard error (SE).

Figure 6 displays the proportion of correct answers for different vertebrate classes in 2006 and 2018. For mammals and amphibians, the results did not show any significant changes over time, but there was a significant difference in the test results concerning the knowledge of birds (−7%, p < .001, d = 0.34), reptiles (−13%, p < .001, d = 0.41) and fish (−30%, p < .001, d = 0.78).

Figure 6. Differences in identification rates of vertebrate classes with the same species tested in 2018 and in 2006 with significant differences indicated by ‘*’.

On the species level, for 12 out of the 20 vertebrates, the results showed a significant change in identification rates. Two of them, the Common Frog and the European Roe deer, could be identified better by pupils in 2018 compared to those in 2006, whereas ten other species were less well known in 2018 (Figure 7).

Figure 7. Changes in identification rates of the different vertebrate species tested in 2018 compared to those in 2006 with significant changes indicated by ‘*’.

The five most and least popular species were the same in 2006 and 2018 (Table 5). Interestingly, two fish species (Northern Pike and Common Carp) lost four and eight places in the ranking of the identification rates, whereas the ‘winners’ (European Roe Deer, Common Frog and Great Spotted Woodpecker) were distributed over three vertebrate classes.

Table 5. Ranking of species’ identification rates in 2006 compared to those in 2018.

Rank	2006	2018
1	Eurasian Red Squirrel	Eurasian Red Squirrel
2	European Mole	European Mole
3	European Badger	Fire salamander +2
4	Common Blackbird	European badger
5	Fire Salamander	Common Blackbird
6	Common Carp	European Roe Deer +5
7	Deaf Adder	European Robin
8	European Robin	Great Spottet Woodpecker +4
9	Northern Pike	Deaf Adder −2
10	Mallard	Mallard
11	European Roe Deer	Common Toad +2
12	Great spottet Woodpecker	Common Frog +5
13	Common Toad	Northern Pike −4
14	Great Tit	Common Carp −8
15	Grass snake	Grass Snake
16	Sand Lizard	Great Tit −2
17	Common Frog	Sand Lizard
18	Common Starling	Common Starling
19	Common Buzzard	Common Buzzard
20	Common Chaffinch	Common Chaffinch

Discussion

Determinants influencing species knowledge

Socio-demographic factors

Gender: Gender has a significant but very small influence on the participants’ species knowledge. This is in contrast to the findings of several studies on Dutch and Swiss schoolchildren (Hooykaas et al., 2019; Remmele & Lindemann-Matthies, 2018) that showed no significant gender differences in their species knowledge. Gerl et al. (2018) and Zahner et al. (2007) reported higher bird species knowledge of girls in German grammar schools. Therefore, gender seems to contribute little to the variance of children’s species knowledge. As (German) professionals in nature conservation and taxonomic sciences or laypeople in environmental clubs are mostly male (Frobel & Schlumprecht, 2016), the key factor affecting boys and girls to pursue their adult professional careers and interests is yet to be discovered.

Size of hometown: Our empirical data do not support the hypothesis that pupils from rural areas have better species knowledge. This is in line with the results of Remmele and Lindemann-Matthies (2018) or other studies that reported greater species knowledge for children of urban areas compared to those of rural areas (Gerl et al., 2018).

A closer look reveals the weak points of the assumption that rural people should outperform all other groups in the test. Many tested animals are difficult to observe in nature because they are either rare, shy, nocturnal or in hidden habitats, that is, their knowledge can hardly be derived from direct observations, but mostly from other information sources (e.g. school, media, family) that are available in urban surroundings as well.

In addition, it is more of a bias than a given fact that the chances of direct nature observations are higher in rural areas. Our data showed that children’s preferences of where to spend their leisure time are the same in rural and urban areas. Furthermore, animals can be spotted more easily in urban parks than in agricultural monocultures because they benefit from a larger variety of ecological niches and more contact with humans (Ossola & Niemelä, 2018). This is in line with Randler et al.’s (2015) study showing that urban park visitors had higher species knowledge compared to non-visitors.

Our finding that the pupils’ school (i.e. their biology teacher) influences their scores in the test are similar to those of Remmele and Lindemann-Matthies (2018), who showed that a mediating person (e.g. a teacher) fosters identification skills. Different teachers might have different abilities in teaching species knowledge and therefore, influence their pupils' test results. This assumption needs further empirical survey because other sociocultural factors (e.g. educational level of parents, leisure time activities of families, etc.) connected with the catchment area of the school might be responsible for the school-related difference in results, too (Tarłowski, 2019).

Mother tongue: Parent–child conversations are important in fostering children’s biological knowledge (e.g. Remmele & Lindemann-Matthies, 2018) but, if done in a different language, children might lack the words to name species in German. Given that many children with a different mother tongue grew up in a different geographic environment with other animal species, it is also possible, that they are not yet familiar with species native to their new home country. Differences in cultural background, that is, the way different cultures look at animals, may also influence species knowledge, too.

A different mother tongue influences educational success in general. Raabe (2019) explained this inequality by the observation that immigrant children are often isolated within their classes, whereas other studies discussed the varying socio-economic status of immigrant and native speaker families as the main cause for the pupils’ different results in standardised tests (Arikan et al., 2017). Further investigation is necessary to clarify whether these lower test results are a linguistic phenomenon or caused by other factors within the social, cultural and economic background of immigrant children. The surprisingly small effect size might be explained by the schooltype we surveyed. A selection process at the end of the fourth grade allows only pupils with excellent language skills to change to grammar schools, so that the sample in our test group was biased with high-performing immigrant children.

Animal observations

Favourite playing area: The results supported our hypothesis (i.e. the more nature observations children make, the higher their species knowledge will be) because pupils who spend more time in nature do have better species knowledge. These findings are similar to those of previous studies (Hooykaas et al., 2019; Randler et al., 2015), but the effect size was surprisingly small for our test group.

Number of excursions: Several studies emphasised the educational potential of first-hand nature experiences (Skarstein & Skarstein, 2020; Waters & Maynard, 2010), but the number of excursions into nature our participants attended during their biology lessons had no significant influence on their species knowledge. This could be explained by the fact that observing animals in nature – especially with a group of children – is rather difficult. Therefore, pupils might acquire their animal species knowledge not mainly by direct observations. Different researchers pointed out the importance of media to pupils’ interest in animals and consequently, their species knowledge (e.g. Prokop et al., 2008) or the role of mediators (e.g. their parents) for fostering species knowledge (Remmele & Lindemann-Matthies, 2018; Tarłowski, 2019). Nevertheless, the small group of pupils in our study who made more than two nature excursions within a school year outperformed all others in the identification test. Although this difference was not significant, it might be worthwhile to repeat this survey with a bigger number of children attending nature excursions more often.

Favourite school subject and personal expectations

Favourite school subject and performance in biology. As the biology curriculum in the tested grammar school classes focused on the human body and general requirements for the survival of vertebrates in an evolutionary context, the tested pupils chose biology as their favourite subject not because they like to identify animals but for other reasons. This might be an explanation of why pupils with biology as their favourite subject did not obtain higher scores than did their classmates who prefer other subjects.

Briede (2016) pointed out that the pedagogic approach of the teacher, but not the content of the subject, is important for pupils’ perceptions of mathematics at school. Assuming that Briede’s findings are generalisable to pupils’ perceptions of other subjects, we believe our findings are not inconsistent with the hypothesis that a high interest in a topic should lead to better scores because our pupils’ choice of their favourite subject does not necessarily reflect their interest in native animals.

The observation that the pupils’ expected performance in biology did not significantly influence their test results might be explained by the different requirements of the tasks in biology lessons and in the species knowledge test. This is because pupils who perform well in biology might also do so likewise in other subjects. In addition, nonformal education processes in families (Remmele & Lindemann-Matthies, 2018; Tarłowski, 2019) might superimpose the effect of formal education on pupils’ species knowledge.

Expected performance in the test: If the participants think they can do well in the identification task, they score significantly higher with a large effect size, but most pupils in our study underestimated their real performance in the test. This might be caused by a rather low self-confidence in identifying animals. To avoid disappointment, children tried to keep their expectations low. These findings indicate an influence of motivational factors on the test results, but further investigation is necessary to examine such influencing factors in detail.

The interpretation of these results in this study seems to suggest that personal expectations influenced the test results as predicted by the expectancy-value theory (Eccles & Wigfield, 2002; Schunk et al., 2014). But the interpretation of the data should be done very carefully because our questionnaire focused on pupils’ performance in this specific test and not on their skills in recognising animals in general. Further investigation should be done to examine the applicability of the expectancy-value theory to predict species knowledge.

Identification rates of vertebrate species

Our result that mammals are the best known animals is in line with other studies on species knowledge of vertebrates (Eschenhagen, 1982; Gerl et al., 2018; Hooykaas et al., 2019; Huxham et al., 2006; Randler, 2006; Remmele & Lindemann-Matthies, 2018). This might be explained by the large number of attractive species of mammals and by their close phylogenetic relationship to humans. As children claim that they like mammals more than all other vertebrate groups (Patrick et al., 2013), it is a logical consequence that mammals are often part of children’s books or television shows to provoke positive emotions. The rich knowledge of these species is therefore rather connected with their omnipresence in media than in personal nature observations. Species that are quite easy to observe such as chaffinches, buzzards or starlings are unknown to most of the tested pupils. Hence, the chance to see an animal is not a good predictor of its popularity.

Pupils are surprisingly unfamiliar with bird species because they are easy to see, colourful and get a lot of attention in the media. Other studies also noticed this special unfamiliarity with birds (Randler, 2006; Remmele & Lindemann-Matthies, 2018; Zahner et al., 2007). Hooykaas et al. (2019) found significant differences in bird identification skills between laypeople (primary school children and general public) and professionals. Professionals had no difficulties in naming birds. The reason for this incongruity between the high visibility and poor identification rates for birds might be the way children look at their natural surroundings. As some birds are small, far away and rather erratic in flight, they all seem to be the same. Children cannot distinguish between different species by linking their special morphological characteristics to their names during the short time of observing them.

Amphibians are the animals in the vertebrate class with the second-highest identification rate because the Fire Salamander is a very attractive species familiar to most pupils. Furthermore, the way we measured species knowledge helped amphibians’ identification rate because pupils scored credits for naming a popular higher taxonomic group such as frogs or toads, even if they did not name the species. This explanation is supported by the big differences between the pupils’ knowledge of the amphibian species and that of a higher taxonomic group. There are two possible explanations for such differences. First, the pupils might have a systematic concept that there are more different frogs than the Common Frog, but they were not sure which frog species it is. Therefore, they named a superordinate taxonomic group as their best answer. Second, the pupils did not have this taxonomic concept. Atran (1998) found that students did not use a taxonomic hierarchy to sort animals. Neither did Polish students, who mentioned several further aspects of how to classify animals besides their morphology (Chyleńska & Rybska, 2018). Therefore, it is very likely, that pupils in our study, who answered ‘frog’ thought that all frogs are the same.

Although fish are difficult to observe due to their submerged life, pupils are quite familiar with them. As all three fish in the test are edible, one explanation for the popularity of fish species is that they knew them as part of their food.

Decline of species knowledge and possible reasons

If the reason for the observed decline in species knowledge is sought in school teaching, one explanation may be curricular changes concerning vertebrates. In 2006, it was compulsory to teach all five vertebrate classes, whereas teachers within the tested school system in 2018 could focus on mammals plus two other classes of their choice. Hence, it is highly probable that today’s pupils encounter fewer vertebrate species in biology lessons than did their peers a decade ago. This might be an explanation for the decline in pupils’ species knowledge over a decade, although no data is available on how many species the pupils really saw during their formal education process.

On the other hand, not only the school’s curriculum has changed but today’s pupils also grow up in a different socio-cultural environment as compared to those back in 2007. Screen time (media-based activities) that were not even available in the early years of the twenty-first century diminishes the time today’s children spent in nature (Larson et al., 2019) and this might be one of the causes for a declining human–nature-connection in general (Colding et al., 2020). Further investigation on the influence of pupils’ screen time and time in nature on their species knowledge will be helpful for an explanation why there is such a decline over a decade.

Conclusion

As a part of environmental knowledge, the ability to recognize species is as important for understanding natural processes as it is for fostering positive attitudes and emotions to conserve nature (Wolff & Skarstein, 2020). Our findings that species knowledge of pupils declined should be regarded as a threat for the environment due to a shifting baseline phenomenon (Leather & Quicke, 2010), but further empirical research with pupils from other educational systems should be done to clarify if the loss of species knowledge is limited to German grammar school pupils or a broader phenomenon.

As environmental education plays an important role in conserving global biodiversity, biology of organisms and taxonomic knowledge should be a focus in the curricula of formal (secondary) education. In addition, efficient teaching strategies to foster species knowledge should be developed. As teachers will be the ones to implement these strategies, their role in passing species knowledge to schoolchildren would be interesting to examine in future surveys. Finally, it could be helpful to foster public species knowledge to define a canon of native animals that every citizen should recognise as part of the general education.

Acknowledgements

The authors would like to thank the pupils and teachers who participated in the research.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

Raw data were generated at the Ludwig-Maximilians University Munich. Derived data supporting the findings of this study are available on request.