Environmental education in zoos: analysis of different scales to measure the impact of educational programs

Abstract

Regular evaluation of environmental education is important for quality assurance and the self-assessment of educational institutions such as zoos. However, capacity is often lacking. Therefore, this study presents instruments that can be used to track the impact of educational programs on connection to nature, interest in species conservation, environmental concern, and attitude toward biodiversity loss. Questionnaires were used before (N_T1 = 674) and after (N_T2 = 619) different educational programs in nine different zoos. Factor analyses were conducted to control for how the scales performed in the zoo context and with visitors of different ages in a pre-post setting. The results showed that the adjusted instruments work well with a few adaptations. Interest in species conservation can be measured as an overall 6-item construct. Biodiversity loss attitude can be measured using a 5-item scale. Using a shortened version of the Environmental Motives Scale, environmental concern was found to be three-dimensional, as intended, and can distinguish effects on egoistic, biospheric, and altruistic concern in zoos. Hence, the scales presented can be used for evaluation purposes and are even applicable beyond zoos.

Keywords:

• Program evaluation
• environmental education
• connection to nature
• environmental concern
• zoo education

1. Introduction

Visiting a zoo is often motivated by the desire to enjoy a day out with friends or family and to have fun (Clayton, Fraser, and Saunders 2009; Coll et al. 2003). The marketing of zoos also tends to focus on the entertainment value (Carr and Cohen 2011). However, zoos have another crucial mission they identify with: education (Patrick et al. 2007; Roe, McConney, and Mansfield 2014). Although Packer and Ballantyne (2002) found that adult aquarium visitors were less interested in the learning aspects of their visit than visitors of museums or art galleries, visitors still associate zoos with education (Roe, McConney, and Mansfield 2014; Ballantyne and Packer 2016). Moreover, learning in the aquarium was perceived as more enjoyable than at the other two sites (Packer and Ballantyne 2002). If zoos are seen as places of fun as well as education, there is an opportunity to combine this day out with environmental messages because education and entertainment are not mutually exclusive, but can complement each other (Packer and Ballantyne 2004). Environmental education is a way to increase environmental awareness and to affect environmental behavior (Liu, Teng, and Han 2020; Varela-Candamio, Novo-Corti, and García-Álvarez 2018). Environmental education is able to positively affect environmental knowledge, attitudes, and intentions (van de Wetering et al. 2022). To this end, zoos offer many different educational programs, such as guided zoo tours, camps and school lessons. Nevertheless, even if education aims to have a positive impact on the visitor, whether it be knowledge, environmental attitudes, or behavior, aiming for an impact is not the same as having an impact (cf. Moss and Esson 2013). Thus, evaluation of environmental education helps to better understand the educational work and to optimize its impact. This is even more important as zoos declare education as an important pillar, as Moss and Esson (2013) point out. For this reason, the European Association of Zoos and Aquaria (EAZA) has issued conservation education standards for its members to follow in their educational work. They state that ‘[t]he zoo must evaluate its conservation education programmes using appropriate methods’ (3). Regarding the effects, a zoo should aim for, they state (EAZA 2016):

Conservation education in zoos should aim to raise awareness of biodiversity loss, connect people to nature and encourage sustainable behaviours (2), and further:

The zoo should aspire to conduct a range of evidence based research to demonstrate the effects of conservation education in zoos has on people’s knowledge, attitude and behaviour towards the natural world (3).

This presents several challenges: Evaluation requires time, statistics require expertise, and capacity is limited. Thus, the aim of the present study is to collect, explain, and evaluate instruments that can then be easily used to assess the impact of zoos and other education institutions on environmental psychological variables inspired by the EAZA education standards. The overarching goal is to be able to use these quantitative instruments in the future to examine whether educational programs lead to changes in connection to nature, interest in species conservation, environmental attitudes, and in the attitude toward the loss of biodiversity.

2. Variables

There is a broad range of approaches to measuring the educational impact of zoo programs or visits on visitors. The variety in published literature shows how differently educational effects are assessed and how diverse they can be. The variables we decided on were chosen to represent the EAZA conservation education standards. This paragraph presents studies related to the chosen variables and explains the idea behind each variable.

2.1. Connection to nature

In general, the term connection to nature is a way to describe the personal identification and relationship with nature (Restall and Conrad 2015), but the definitions differ slightly. Often cited is Aldo Leopold, who wrote in 1948: ‘When we see land as a community to which we belong, we may begin to use it with love and respect’ (in Leopold and Finch 1987, viii). This reflects the basic idea well: Higher connection to nature is associated with enhanced environmentally friendly behavior (Pereira and Forster 2015; Geng et al. 2015; Mayer and Frantz 2004). In contrast, alienation through increasing physical and psychological separation from nature can be seen as a possible cause of many environmental problems (Schultz 2002; Zylstra et al. 2014). Besides environmental behavior, connection to nature is also positively correlated with various forms of well-being, such as psychological well-being, happiness, and self-rated health (Capaldi, Dopko, and Zelenski 2014; Howell et al. 2011; Nisbet and Zelenski 2013; Puhakka et al. 2018).

Studies show that connection to nature is associated with experiences and time spent in nature (Braun and Dierkes 2017, 2017; Nisbet, Zelenski, and Murphy 2009; Schultz and Tabanico 2007). Moreover, noticing nature is a possible way to increase connection to nature (Richardson et al. 2022). That makes zoos ideal places to promote this connection and work against alienation from nature. Bruni, Fraser, and Schultz (2008) found that a zoo visit significantly increased visitors’ implicit connection to nature, as measured indirectly by using a program that required participants to sort words into two categories, but did not increase explicit connection to nature, as measured directly by using a questionnaire. In contrast, Kleespies et al. (2020, 2022) findings showed guided zoo tours were able to significantly increase explicit connection to nature in high-school students with low to moderate connection to nature as well as in adults with already high connection to nature. They concluded that guided tours can be an effective way to increase connection to nature. According to Frantz and Mayer (2014), increasing people’s connection to nature should be an important goal of environmental education. Furthermore, EAZA expects that its member zoos strive to connect people to nature. As such, connection to nature is a valuable variable that zoos should consider evaluating on a regular basis.

2.2. Interest in species conservation

The EAZA conservation education standards suggest that education should aim to increase knowledge, but do not specify the subject of knowledge. A comparative literature review by Schilbert and Scheersoi (2023) shows that knowledge-related outcomes are the most commonly measured cognitive dimension in zoo studies. For example, Jensen (2014) and Wagoner and Jensen (2010) used animal drawings made by children and teenagers to demonstrate that visiting a zoo or participating in a program can lead to new knowledge about animals and their habitats, as well as correct misconceptions. Guided visits were more effective than unguided visits (Jensen 2014). The goal of our study is to present tools that can be used across zoos, thus for our purposes we chose to measure interest rather than knowledge to be independent of specific program content. Interest is often considered to be related to motivation and learning (Hidi 2006; Hidi and Harackiewicz 2000; Schiefele 1991). As a motivational variable, interest refers to engaging or wanting to engage in the future with specific content, i.e. objects, ideas, and events (Hidi and Renninger 2006). It describes a person-object relationship with an object from the individual’s environment, usually involving positive emotions toward the object (Krapp 1993). There are two types of interest (Ainley 2017; Hidi 1990; Krapp, Hidi, and Renninger 1992): Situational interest is evoked by something in the environment, is more focused on the present moment and mainly short-term. Individual interest (personal interest) is the tendency or predisposition of the individual toward a specific content, focusing more on the person and less on the situation, and is therefore longer lasting and evolves over time. Our focus is individual interest (Kleespies, Doderer, et al. 2021). Individual interest can evolve out of repeated situational interest (Hidi and Renninger 2006; Palmer, Dixon, and Archer 2017; Rotgans and Schmidt 2017a). Rotgans and Schmidt (2018) argue that the impact of individual interest on knowledge gain is mediated by situational interest. Individual interest is a possible consequence of knowledge acquisition (Rotgans and Schmidt 2017b).

To our knowledge, there are not as many studies on interest as on knowledge in the zoo context. Moss and Esson (2010) measured interest by observing the time spent in front of an enclosure and came to the conclusion that especially the taxonomic group, but also body size and activity by the animal are decisive for interest in an animal. Furthermore, Dohn (2011) found that an excursion to an aquarium generates situational interest. As species conservation is a crucial topic for zoos (cf. Patrick et al. 2007; Barongi et al. 2015), our focus is interest in species conservation.

2.3. Environmental concern

According to EAZA’s standards, conservation education in zoos should positively influence attitudes toward the natural world. For the purpose of this study environmental attitudes are defined according to Gifford and Sussman (2012) as ‘concern for the environment’ and the ‘caring about environmental issues’ (65). With the goal to implement scales to measure attitude in zoos, we assess two types of attitudes: first, the general attitude toward the environment in form of this environmental concern, and second, the specific attitude toward biodiversity loss (see section 2.4).

Various authors have already studied different types of attitudes in zoos. Hacker and Miller (2016) assessed the impact of the elephant enclosure at San Diego zoo on visitors and found visitors who had a close-up encounter and observed more active behavior in elephants had a stronger conservation intention, while a talk with zoo staff had no significant effect. Lukas and Ross (2005) found no change in attitudes toward gorillas and chimpanzees following a visit to the ape facility, but a difference in attitudes for gorillas and chimpanzees. The attitudes toward the apes were mostly naturalistic and moralistic, but the attitude toward chimpanzees was more dominionistic and the attitude toward gorillas more negativistic. More frequent visitors held a higher ecoscientistic attitude. Martens, Hansart, and Su (2019) reported lower animal attitudes, meaning less concern and respect for animals, among high school students who had never visited a zoo or aquarium compared to those who had visited more often. In a study by Moss and Pavitt (2019), there was a change in conservation attitudes after a walk-through exhibit, which might be also due to the whole zoo experience. Kleespies, Montes, et al. (2021) analyzed which factors influence the attitude toward species conservation in zoo visitors, and found age, frequency of zoo visits, perception of zoos, interest in animals, and country as influential factors.

2.4. Biodiversity loss attitude

Besides climate change, biodiversity loss is a pressing problem that has not only ecological but also social consequences. Biodiversity and ecosystems influence each other, which means that ecosystem services, the benefits humans derive from ecosystems (Millennium Ecosystem Assessment 2003), and ecosystem functioning are also affected by a loss of biodiversity (Cardinale et al. 2012; Chapin et al. 2000). Currently, biodiversity is under threat and declining at an unusually high rate, unprecedented in human history (Ceballos et al. 2015). There are many direct and indirect ways in which humans impact nature and biodiversity (Balvanera et al. 2019). The most urgent drivers of biodiversity loss are land-use change and exploitation (Caro et al. 2022), but also climate change plays an important role (Sala et al. 2000). In 2010, the United Nations developed a framework to address the issue of biodiversity loss. Four goals were developed to address this issue, based on 20 targets. Target No.1 states that by 2020 everyone should be aware of the value of biodiversity and how to conserve it (Secretariat of the Convention on Biological Diversity & United Nations Environment Programme 2019). Zoos, with their diversity of animal life, are ideal places to experience biodiversity. Moss, Jensen, and Gusset (2015) conducted a study that showed that understanding of biodiversity improved after a visit to the zoo. Such effects can also last over two years and knowledge about biodiversity measures may even increase (Jensen, Moss, and Gusset 2017). In Colléony et al.’s research (Colléony et al. 2016), zoo visitors were significantly more concerned about threats to biodiversity than park visitors, and this concern was positively related to the number of zoo visits, but the visit itself did not affect the concern. On the other hand, in a study by Clayton et al. (2017), a zoo visit slightly influenced concern about threats to biodiversity. In our study, biodiversity loss attitude is also measured in terms of concern about the loss but with a different scale.

3. Methods

3.1. Questionnaire and procedure

Our study was carried out in nine zoos using questionnaires. The participants of educational programs (e.g. a guided tour) filled in a questionnaire with 25 questions at the beginning of the program (T1) and immediately after the program (T2, 22 questions). T1 included additional demographic information and frequency of zoo visits. The participating zoos conducted their educational programs as usual, with the extra of the questionnaires, which were completed on site. The questionnaire was anonymous and voluntary. This study used an individual code for each participant to track the change due to the program. The study was approved by an ethics committee of the lead authors’ university. For participants under the age of 18, the written consent of a parent or legal guardian has been obtained. There was no disadvantage for non-participation. Data was contributed by seven zoos in Germany, one in Switzerland (both in German language), and one in the Netherlands (in Dutch), from mid-July 2022 to mid-July 2023.

3.2. Participants

The study participants were zoo visitors who participated in a program that was either booked or public, such as company outings, family trips, school trips, and others. Overall, 674 T1-questionnaires and 619 T2-questionnaires were completed. For a more precise analysis, age groups and language were taken into account and the sample was divided into Dutch (hereinafter: NL sample) and German language (Germany, Switzerland), whereby the German/Swiss sample was divided into children (<18 years, N_T1= 296) and adults (N_T1= 268). As they did not indicate their age, 11 people were excluded. Further, T2 questionnaires without matching T1 questionnaires were excluded because age was not known (10 questionnaires). The German/Swiss children sample has a mean age of 14.26 years and a median of 15 years (47.0% male; 49.3% female; 2.4% diverse; 1.4% no answer). For the sample of adults, the mean age is 42.01 years with a median of 41 years (22.0% male, 73.9% female; 0.7% diverse; 3.4% no answer). The NL sample included (except for one person) only adults (N_T1= 99).

The NL sample (T1) has an average age of 49.88 years, median of 52, with one person below 18 years (17 years old) and the rest being adult (42.3% male; 43.2% female; 0.9% diverse; 1.8% no answer).

Since not all items were fully answered, the sample size (N) of the valid questionnaires used for calculations varies between the different scales. The respective N on which the calculations are based are shown in Table 1.

Table 1. Overview of sample sizes (N) in each category. N_total indicates the number of questionnaires, n_valid indicates the sample size after listwise deletion (per scale).

		Adults		Children		NL
		T1	T2	T1	T2	T1	T2
Ntotal		268	237	296	263	99	101
Nvalid	connection to nature	266	234	294	251	99	101
	interest in species conservation	259	233	289	261	94	100
	biospheric subscale	259	235	286	258	99	101
	egoistic subscale	258	233	287	258	97	100
	altruistic subscale	260	235	289	258	98	101
	biodiversity loss attitute	263	233	286	256	98	101

3.3. Measurements

3.3.1. Factor analyses

Exploratory factor analysis in form of Principal Axis Factoring (PAF) was used to test the instruments for underlying dimensions since all scales were adjusted or at least shortened (except IINS). An oblique rotation method (oblimin) was chosen as correlations between factors can be assumed (Costello and Osborne 2005). If a participant did not complete all items on a scale, they were excluded from factor analysis of that same scale (listwise exclusion criterion).

PAF was performed on T1 and T2 separately, both for children and adults. For the Dutch Zoo, a separate PAF was performed. Thus, the factor structure for each scale was tested six times (adults T1/T2, children T1/T2, and NL sample T1/T2). This was done to ensure that the instruments are usable for different ages, languages, and in a pre-post comparison. Further, the Kaiser–Meyer–Olkin (KMO) test for sampling adequacy and the Bartlett’s test of sphericity were performed to assure factor analysis is appropriate for the data (Field 2018). Cronbach’s α was calculated for each resulting factor to check reliability (Tavakol and Dennick 2011).

3.3.2. Correlations

The Shapiro-Wilk test was used testing for normal distribution; this was supported by visual analysis of histograms (not shown). The factors of the different data sets were not normally distributed (p < 0.001). Therefore, Kendall’s τ was chosen to measure correlation between the instruments and factors. Correlations were calculated for T1 for adults and children to determine general relationships.

4. Instruments

In the following, the different instruments used to measure the variables introduced in the upper sections (2.1 to 2.4) are presented. The scales have been chosen to reasonably reflect the EAZA standards as well as to be relatively short in order to facilitate their use in zoos.

4.1. Connection to nature

There are various scales for measuring connection to nature, but we decided on the IINS (Kleespies, Braun, et al. 2021), a graphical extension of the Inclusion of Nature in Self (INS) scale by Schultz (2002). In this definition of connection to nature, Schultz (2002) focuses on cognitive connection, i.e. the extent to which ‘an individual includes nature within his/her cognitive representation of self’ (67). The participant is asked which graphic best describes their relationship with nature. There are seven pairs of two circles to choose from. The circles are overlapping to varying degrees, one stating ‘self’ and the other ‘nature’. ‘Nature’ is represented by several characteristic elements often associated with nature (Kleespies, Braun, et al. 2021) and the ‘self’ by the silhouette of a person. This graphical variant was chosen to make the tool more accessible, especially to children. Since this scale has only one item, no factor analysis was performed.

4.2. Interest in species conservation

The chosen instrument, the Nature Interest Scale (NIS) (Kleespies, Doderer, et al. 2021), differentiates between three dimensions: emotional, cognitive, and value-related interest, which we found quite suitable for zoo context. It could show, in which way (emotional, cognitive, and/or value-based) the people are affected by the zoo program. The word ‘nature’ was exchanged for ‘species conservation’. The scale was shortened to two items per dimension (see Table 3). The statements are answered using a 5-point Likert agreement scale (from totally disagree to totally agree).

Table 3. Items of the interest in species conservation scale and factor loadings for each subsample.

			Adults		Children
		Originally intended as^a	T1	T2	T1	T2
1	I would like to know much more about species conservation.	cognitive	0.914	.912	.867	.902
2	I would like to learn more about species conservation.	cognitive	.896	.881	.867	.879
3	I find it exciting to deal with species conservation.	emotional	.868	.898	.829	.857
4	Learning about species conservation is fun for me.	emotional	.817	.891	.831	.882
5	I find it meaningful to be involved with species conservation.	value-related	.801	.751	.673	.807
6	I think it’s important that I am well informed about species conservation.^b^‡	value-related	.784	.788	.625	.778

^a As the scale is considered unidimensional, there are no factors other than the overall construct of interest.

^b The English item from Kleespies, Doderer, et al. (2021) has been slightly reworded here to be more in line with the German translation.

4.3. Environmental concern

There are various popular scales for measuring different kinds of environmental attitudes (for an overview, see Cruz and Manata 2020 and McIntyre and Milfont 2016). Cruz and Manata (2020) recommend the use of the Environmental Motives Scale by Schultz (2000, 2001) because it is short and at the same time has good reliability. Therefore, and because it is an easily comprehensible scale fitting the zoo context, it is also used here for measuring environmental concern. The scale is based on Stern and Dietz (1994) value-basis theory for environmental attitudes. According to their relation to the self, the relative value a person places on self, others, and plants and animals is considered (Schultz 2000). The scale states ‘I am concerned about environmental problems because of the consequences for…’. Each item can be rated from not important to supreme important. Three dimensions are covered: biospheric, altruistic, and egoistic concerns about environmental problems. The original scale has twelve items, four of each component. It was shortened to three items per component so that the questionnaire would not be too long. The original scale is a 7-point Likert scale, but since two other scales use 5-point scales, for consistency, a 5-point scale was preferred. The items can be found in Table 4.

Table 4. Items of the Environmental Motives Scale by Schultz (2001) (shortened) and factor loadings (pattern matrix) for adults and children (T1, T2). Values below 0.3 are not shown.

		Adults						Children
		T1			T2			T1			T2
		Altruistic	Biospheric	Egoistic	Altruistic	Biospheric	Egoistic	Altruistic	Biospheric	Egoistic	Altruistic	Biospheric	Egoistic
1	Children	.931			.846			.903			−0.989
2	All people	.792			.756			.795			−0.609
3	Future generations	.714			.851			.534			−0.485
4	Marine life		−0.913			−0.979			.694			.883
5	Animals		−0.865			−0.874			.898			.911
6	Plants		−0.599			−0.714			.500			.811
7	My future			.764			.832			.767			.692
8	My lifestyle			.763			.754			.770			.877
9	My health			.747			.786	.340		.546			.697

4.4. Biodiversity loss attitude

For measuring attitude toward biodiversity loss, the Climate Change Attitude Scale (CCAS) by Christensen and Knezek (2015) was adapted for this new topic. It was developed and tested for middle school students and the usage in a pre-post setting is proposed, which is quite fitting for the aim of our study. The scale consists of two factors: beliefs and intentions. The term ‘climate change’ was exchanged by ‘loss of biological diversity’. The next step was shortening the scale for application in the zoo. The items of the intentions sub-scale are more related to general environmental issues, so we focused on the belief-factor. We only adapted one item (item 6, Table 6) from the intentions factor. This reduced the scale to six items. The participants responded to the statements on a 5-point agreement scale.

Table 6. Items of the biodiversity loss attitude scale (inspired by Christensen and Knezek 2015 and factor loadings (pattern matrix) for each subsample.

		Adults		Children
		T1	T2	T1	T2
1	I am concerned about the loss of biological diversity.	.717	.768	.648	.745
2	The loss of biological diversity will impact future generations.	.788	.827	.712	.780
3	The loss of biological diversity has a negative effect on our lives.	.789	.744	.715	.793
4	The loss of biological diversity will impact our environment in the next 10 years.	.739	.589	.732	.829
5	Human activities cause the loss of biological diversity.	.579	.752	.704	.712
6	We cannot do anything to stop the loss of biological diversity.	excluded	excluded	excluded	excluded

5. Results

According to the KMO and the Bartlett test of sphericity, all six data sets are applicable for factor analysis. Bartlett’s test was significant for each scale and group (p < 0.001) and KMO values were always higher than 0.7, ranging from middling to excellent (Kaiser and Rice 1974). Detailed KMO values can be found in the appendix (Table 1). The results of the German-language samples are described first, followed by a short section on the NL sample.

5.1. Interest in species conservation

The PAF showed that the most reasonable solution is to consider interest as single-factorial for adults and children, for T1 as well as for T2. This is supported by both eigenvalues (Table 2) and scree plots (Figure S1). The expected three dimensions (cognitive, emotional, and value-related interest) do not emerge even when fixing the analysis to three factors. Therefore, interest in species conservation is here considered to be unidimensional and a solution with three dimensions is rejected. Cronbach’s α values for this unidimensional solution are greater than 0.9 and therefore acceptable (Peterson 1994) (adults: T1 α = 0.938, T2 α = 0.940; children: T1 α = 0.905, T2 α = 0.940). The factor loadings are given in Table 3.

Table 2. Results of factor analysis. The three highest eigenvalues for children and adults, T1 and T2, of interest in species conservation, biodiversity loss attitude, and environmental concern, and the variance (var.) they each explain.

Adults				Children
T1	var. [%]	T2	var. [%]	T1	var. [%]	T2	var. [%]
Interest in species conservation
4.587	76.45	4.646	77.43	4.076	67.93	4.624	77.06
.483	8.05	.467	7.78	.594	9.90	.416	6.93
.322	5.37	.365	6.08	.523	8.71	.320	5.33
Environmental concern
4.377	48.63	4.388	48.76	4.517	50.18	4.949	54.99
1.523	16.92	1.752	19.47	1.465	16.28	1.686	18.73
1.055	11.73	1.105	12.28	.747	8.30	.693	7.70
Biodiversity loss attitude
3.094	61.89	3.174	63.49	2.973	59.46	3.384	67.68
.650	12.99	.671	13.41	.611	12.23	.512	10.23
.497	9.93	.426	8.52	.550	11.00	.429	8.58

5.2. Environmental concern

5.2.1. Adults

Factor analysis revealed three factors with eigenvalues above 1 (Kaiser-criterion) for T1 (see Table 2). The scree plot allows two interpretations: one- or three-factorial (Figure S2). Based on the theoretical consideration that environmental concern consists of three dimensions, the three-factor solution was chosen, which then represents the three dimensions (biospheric, egoistic, and altruistic) well (Table 4). A very similar picture emerged for T2. Again, the three-factor solution appears to be reasonable, the three components are represented as originally intended. The Cronbach’s α values can be categorized as ‘good’ (see Table 5).

Table 5. Cronbach’s α for the three factors biospheric, egoistic, and altruistic concern in each subsample.

Factor	Biospheric				Egoistic				Altruistic
Group	Adults		Children		Adults		Children		Adults		Children
	T1	T2	T1	T2	T1	T2	T1	T2	T1	T2	T1	T2
α	.852	.888	.787	.902	.805	.838	.801	.855	.868	.871	.865	.876

5.2.2. Children

Based on Kaiser-criterion, there are two factors (Table 2), but since the theoretical basis is three dimensions, factor analysis was fixed to three factors in the next step. This resulted in the expected three-factorial solution of biospheric, altruistic, and egoistic concern as seen for adults and as justified by Schultz’s (2001) theory. Cronbach’s α values range from acceptable to good (Table 5).

5.3. Biodiversity loss attitude

Based on eigenvalues, the PAF showed two factors for adults in T1 and T2. The item that was part of the intentions sub-scale (item 6, Table 6) loads highest on a second factor, but with a loading below .3. For the children sample, only one factor emerges, but the same item has loadings below an absolute value of .3. Further, Cronbach’s α improves considerably when item 6 is deleted.

With these considerations, this item was excluded from analysis, factor analysis was repeated and with the new results, this scale is considered as unidimensional (eigenvalues in Table 2). Factor loadings are given for this 5-item solution in Table 6 and scree plots in app. Figure S3. Cronbach’s α scores are thus: adults: T1 α = 0.839, T2 α = 0.848; children: T1 α = 0.829, and T2 α = 0.879.

5.4. NL sample

The results for the NL sample are close to those of the German language samples. They fulfill the criteria for factor analysis and similar factors are showing, despite the different language. Again, for the interest scale and also for the biodiversity loss attitude scale, the single-factorial solution is found to work best, whereby the exclusion of the one intention item of the biodiversity scale also proves here to be necessary. At the same time, for environmental concern the three-factorial solution works well. Detailed analysis can be found in the appendix (Tables 2–5).

5.5. Correlations between (sub)scales

Kendall’s τ shows that there are significant positive correlations between all scales. Overall, in the adults and in the children sample, altruistic and biospheric concern as well as altruistic and egoistic concern correlate the highest. Even if the egoistic subscale correlates significantly with interest in species conservation, biodiversity loss attitude, and the IINS, it only correlates weakly (τ < 0.2). (Tables 7 and 8). In contrast, the biospheric subscale correlates well with those three other scales (τ > 0.3).

Table 7. Correlations between the different (sub)scales for the adults T1 sample: coefficient (above), and sample size (pairwise; below).

T1/adults		2	3	4	5	IINS
1 Interest in species conservation	coef.	.355**	.367**	.098*	.172**	.326**
	n	258	254	253	255	257
2 Biodiversity loss attitude	coef.		.374**	.155**	.293**	.201**
	n		258	257	259	261
3 Biospheric concern	coef.			.336**	.453**	.265**
	n			254	256	257
4 Egoistic concern	coef.				.445**	.153**
	n				256	256
5 Altruistic concern	coef.					0.168**
	n					258

The asterisks indicate significance: *p < 0.05; **p < 0.01.

Table 8. Correlations between the different (sub)scales for the children T1 sample: coefficient (above), and sample size (pairwise; below).

T1/children		2	3	4	5	IINS
1 Interest in species conservation	coef.	.340**	.328**	.113**	.276**	.325**
	n	284	283	285	286	287
2 Biodiversity loss attitude	coef.		.375**	.100*	.247**	.228**
	n		281	283	284	284
3 Biospheric concern	coef.			.279**	.476**	.376**
	n			283	284	285
4 Egoistic concern	coef.				.510**	.164**
	n				285	285
5 Altruistic concern	coef.					.263**
	n					287

The asterisks indicate significance: *p < 0.05; **p < 0.01.

6. Discussion and implications

The purposes of this paper were to present and analyse scales that can be used for evaluation in zoos and other environmental education institutions. The variables and instruments were applied in various zoos and analyzed for their factor structure.

The scale of interest in species conservation did not show the expected factors of different interest types, and is therefore considered unidimensional. This means, when the instrument is applied in a zoo, this instrument is able to measure if there is a change in the overall interest in species conservation, but it is not possible to distinguish between a change in e.g. emotional or cognitive interest. For biodiversity loss attitude, two factors were expected, but PFA showed a single factor solution is more appropriate. Hence, it is best measured as a 5-item construct of overall biodiversity concern. For environmental concern, the expected three factors emerged. In other words, this tool can be used to distinguish the impact of an educational program on egoistic, altruistic, and biospheric concerns. This is interesting because these concerns are differently related to environmental behavior. Positive associations between altruistic and biospheric concern and pro-environmental behavior are known (Pereira und Forster 2015; Milfont, Duckitt, and Cameron 2006; Schultz et al. 2004), but this might depend on the group studied. Milfont, Duckitt, and Cameron (2006) found for Asian New Zealanders a positive association between altruistic and biological concern and environmental behavior, but for European New Zealanders only between biospheric concern and behavior. Likewise, Schultz et al. (2004) found no correlation between altruistic concern and environmental behavior. Moreover, Schultz et al. (2004) and Schultz et al. (2005) even described negative associations between egoistic concern and pro-environmental behavior. At the same time, De Dominicis, Schultz, and Bonaiuto (2017) show that egoistic motives (self-interest) still can cause environmental behavior. It is beneficial when zoos succeed in increasing concern about the environment. Although biospheric, altruistic, and egoistic concerns were positively correlated in our study, based on the known studies, it is better to rely on altruistic and biospheric concern, as the results regarding a positive influence on environmental behavior are clearer.

All instruments and subscales correlated positively. This is also true for the egoistic subscale, even if it had the weakest correlations. This is in contrast to Mayer and Frantz (2004) and Schultz (2001), who found no significant correlations for egoistic concern and connection to nature. Correlation in general cannot prove a causal relationship, but Kendall’s τ does imply a positive monotonic relationship (Mirtagioglu and Mendeş 2022). A high T1 score on one scale of our study indicates a likely high score on another (and vice versa). This helps to understand the different types of participants in zoo programs. A person who comes to the zoo with a strong connection to nature is also likely to have a greater interest in species conservation and an increased concern for the environment. Similar to the findings of Yocco et al. (2015), altruistic concern correlated more strongly with the biospheric and egoistic factors than the latter two with each other. Among environmental concern, the biospheric factor was highest correlated with connection to nature, which is in line with other authors (Mayer and Frantz 2004; Schultz 2001). In addition to biospheric concern, the connection to nature is of particular interest when considering environmental behavior. If zoos can increase this connection, they can help build a bridge to nature and help reduce a much-discussed cause of environmental problems, the alienation from nature, by reconnecting people to nature (cf. Ives et al. 2018).

We aimed to implement scales to measure the impact of education on connection to nature, interest in species conservation, environmental concern, and biodiversity loss attitude. Regarding the usage of the proposed scales, to our knowledge, the scales of interest in species conservation and biodiversity loss attitude have not been used in a zoo context before. But there are studies resembling those ideas: Kleespies, Montes, et al. (2021) found interest in animals an important factor for attitude toward species conservation. Additionally, Clayton et al. (2017) measured concern about threats to biodiversity with a single-item scale - the zoo visit slightly affected the concern. The Environmental Motives Scale for environmental concern was already used in zoo-context by Yocco et al. (2015). They found that zoo visitors have comparatively high biospheric concern (whereby socio-altruistic concern was the highest) and also agreed most to biospheric messages, whereby the frequency of zoo visits and membership played a role. The change in connection to nature due to participating in a zoo program was researched by Kleespies et al. (2020, 2022). For evaluation purposes it is important to learn whether there are changes (positive or negative trends) as a result of a zoo’s educational program, to see, in what variables there is a change, and possibly to compare any differences between programs. For evaluation, a test of significance can be performed using the mean values of T1 and T2 for each (sub)scale. With this, the impact of an educational program can be measured.

As can be seen from the many zoo studies existing, there are many different aspects to be studied. There is no one single variable that gives perfect feedback on zoo education, but there are many aspects interesting for research. The variables analyzed were inspired by the EAZA conservation education standards, so the variables can be considered relevant for zoos and related educational institutions. We selected aspects we believe are universal elements of zoo education, are important issues of our time, and may help reflect the messages zoos are trying to send. While a declining connection to nature is considered a cause of environmental problems, interest helps people to be more attentive, and if they care about environmental problems such as biodiversity loss, this could increase support for conservation efforts. The quantitative survey method allowed a large number of zoos to participate without the assistance of an additional researcher. By that, a comprehensive data set could be created – however, a disadvantage of quantitative methods is that they might miss effects that qualitative methods could show, leaving these effects unidentified (Moss and Esson 2013). The purpose of our study was to analyze how the instruments behave under different conditions. We analyzed the data divided by age and language (whereby a special characteristic of the NL sample is that most of its participants were more frequent visitors to the zoo) and come to the conclusion that the instruments are well suited for evaluating zoo education programs. Overall, the factor structures proved to be stable, so before-and-after comparisons are possible using these instruments. The proposed variables have the advantage of not being dependent on elements such as specific enclosures. Therefore, instead of knowledge we propose to measure interest and for the same reason, we propose the use of environmental concern. We hope that the explanations of the variables and instruments presented here can be used as a toolbox and starting point for evaluation efforts.

The Conservation Education Standards have been updated in 2023, in which is stated that (EAZA) zoos and aquariums are ‘a valuable resource for conservation education and meeting many of the United Nations Sustainable Development Goals’ (EAZA 2023, 1). Zoos are well-positioned to combine a potential day of recreation with environmental messaging and education. However, this should not be done blindly, and in addition to external monitoring, it is important that zoos engage in self-assessment and reflection. Conducting studies and evaluating the impacts of zoo programs could improve the way to think about the educational role. In the long term, education for sustainable development in its various forms should also reach those visitors who are not actively participating in a program and/or have no initial interest in education on their zoo day. Looking at the potential that zoos have in their programs helps to understand what might be possible.

7. Limitations

As in every study, some limitations have to be taken into account. Regarding the procedure, many zoos participated, meaning, there were many people (like zoo guides) associated with this project, leading to possibly slight differences in the handling of the questionnaires. Some (a small proportion) questionnaires were black-white copies. Since the survey was voluntarily, only those who were willing to complete the questionnaire are included in the data, which can lead to bias. Methodologically, terms such as ‘biological diversity’ or ‘species conservation’ may present some difficulties, as there may be differences in understanding between individuals. For example, biodiversity is often considered abstract (Consorte-McCrea et al. 2017). A recent study showed how wide-ranging associations with biodiversity and its loss may be (Eylering et al. 2023). Further, our study comprises multiple program types. For the purpose of our study, they were combined into a larger data set and factor analyses were the focus. Especially for T2 though, the content of the program may influence the factor structure.

CRediT author statement

V. Feucht: Conceptualization; Methodology; Formal analysis; Data curation; Writing – Original Draft; Project administration; P. W. Dierkes: Conceptualization; Writing – Review & Editing; Supervision; Funding acquisition; M. W. Kleespies: Conceptualization; Writing – Review & Editing; Supervision; A. Haeser-Kalthoff, A. Kästner, C. Mager, S. Reichhardt, L. Schröder, T. Spengler, P. Ugolini, F. Waked: Investigation.

Ethics statement

This study was reviewed and approved by the Ethics Committee of the Science Didactic Institutes and Departments (FB 13, 14, 15) of the Goethe University Frankfurt am Main (Reference number: 15-WeWiLuHa-2205).

Acknowledgements

We would particularly like to thank all the zoos and the staff that supported us in carrying out the study: Grüner Zoo Wuppertal, Koninklijke Burgers’ Zoo, Neunkircher Zoo, Opel-Zoo Kronberg, Tierpark Hagenbeck, Zoo Basel, Zoologischer Garten Köln, Zoo Leipzig, and Zoo Neuwied.

Disclosure statement

Member zoos of VdZ participated in the collection of data. The authors declare no conflict of interest. The funders had no influence on study design and analysis.

Additional information

Funding

This study was partly supported by the Opel-Zoo foundation professorship in zoo biology from the ‘von Opel Hessische Zoostiftung’ and the Verband der Zoologischen Gärten (VdZ).