Recommendation 1: Start with listening
If science communication activities are not to be planned solely based on assumptions and stereotypes, a reflection on one's activities and goals is necessary. Above all, precise analysis and knowledge of the audience reached and not reached so far is required. For this, it is important to listen and ask questions first: How do others perceive science, science communication, or even individual research topics? Which needs do they express? What do they wish for? Even if the answers are not formulated ready for implementation, open discussions can bring many insights. It is important not only to listen but also to respect these wishes. Especially, it is important to respectfully accept a “no” for an answer. It can happen that initially there is no interest in a certain topic within a community, although this might have high relevance for people. For example, if people are facing acute financial challenges, a discussion event on the importance of future technologies in 10 years might be of little direct relevance to them.
Do not expect people to change and become “like you”. Rather, “the intention must be to seek out and embrace, on their own terms, the ingenuities that continually arise in the shadows or as subversions of the established narratives” (Coffee, 2008, p. 271; cf. Archer et al., 2016).
This implies that you have to be able to adapt your plans to the needs and wishes of the communities and self-critically reflect the contents you want to communicate and how they are communicated. This process of (self-) reflection and adaptation should be done regularly throughout the particular project (Aguirre, 2014, p. 11; Marschalek and Schrammel, 2017, p. 25–26).

Recommendation 2: Reduce the Distance and Be Accessible
[D]istance can be expressed on many levels: be it as an academic or upper-class language with respect to vocabulary, idioms, or references, a condescending and instructive attitude, or the display of and insisting on academic titles—in short, a certain habitus (Bourdieu, 1982). Since this distance is often the result of one's own—especially socioeconomic—living conditions and the educational system, it will not be possible for a science communication project to overcome it easily. But there are strategies to at least narrow the gap. For example, you could formulate language in a casual, humorous, and colloquial way. However, at the same time you have to remain authentic and not play an artificial role. (…)
The dialogue should take place at eye level. Our experience shows that (scientific) expertise is respected—titles and references to organizations alone not necessarily: Initially, the aim should be to sound out emotions, attitudes, and values to create a common basis on which further discussion and knowledge transfer are possible. Current socio-psychological studies on values and emotions in science show that certain attitudes are driven by values and emotions to such an extent that one cannot achieve anything with information events alone. Dan Kahan, for example, described this observation, which is discussed under the heading “Cultural Cognition” (Kahan et al., 2010b), for information on vaccination (Kahan et al., 2010a).
To this end, it is “necessary to honestly question one's own goals and how they could be achieved” (Marschalek and Schrammel, 2017, p. 26) in advance. How should results be dealt with? Is there even a need for concrete results or a comprehensive transfer of knowledge, or is the exchange itself a goal? (…)

Recommendation 3: Be Relevant for Everyday Life
Science in general and thus also science communication seemed to them to be remote from their everyday life, inaccessible, complicated and correspondingly uninteresting. For this reason, concrete topics or hooks that tie in with already existing interests or life situations (e.g. job related links, or leisure time activity) contribute decisively to the success of science communication.
Establishing these links “between their homes, personal lives, communities, and science are important” (Archer et al., 2016, p. 936), as experiences from other projects show, too. For example, Marschalek and Schrammel (2017, p.28), state that in their project “exhibition objects or exhibition themes with a relation to the everyday life of the target group, create particular interest and encourage coming back” (cf. Streicher et al., 2014). 
The same holds for connections to the cultural background and other experiences of the underserved audiences (Archer et al., 2016, p. 936).
The relevance of topics cannot be measured solely by whether they relate to everyday life in terms of content. Starting points can also be found in more pragmatic aspects not related to the topic, for example in a scientific holiday program for children, which offers free care as a benefit, or in an entertaining scientific event that offers an interesting leisure activity away from daily struggles. In many cases we have noticed that science hardly plays a role outside very concrete benefit considerations—for example, to support career plans or as education for their children. This is usually not a rejection, but a low priority compared to other topics (see rec. 1).

Recommendation 4: Go Where People Are
If you want to address groups that have not been reached so far, it is helpful to also approach these groups quite literally in a spatial sense. You should use places and buildings that are familiar, easily reachable and accessible for the group—“localities in the everyday environment of the people” (Marschalek and Schrammel, 2017, p. 22), because “socially inclusive science communication has to take place where people spend most of their time—within their communities” (Streicher et al., 2014, p. 1). (…) 
In concrete terms, this means organizing events in the district center, the village pub, or vocational school instead of the university or research institute—which often are perceived as closed-off areas. (...)
In addition to closeness and physical accessibility, (...) there are places some people just won't go because they feel that these places are “not for them” (Dawson, 2019, p. 100–102). Specifically for science communication, both a fenced research center with guards and admission controls and a classicist science building (which already seems to signal on the façade that you can't get very far here without being able to have fluent conversations in Latin) for example do not appear welcoming to many people (Marschalek and Schrammel, 2017, p. 22). Thus, it seems important to mind easy accessibility and reachability of the location, as well as a familiar and open atmosphere, a direct dialogue with the community and multilingual information and offerings (Archer et al., 2016, p. 936; Streicher et al., 2014). This so-called outreach approach can also mean connecting to existing events with a communication format, for example being represented with a booth at a block party or giving a lecture at an event of a local initiative or an association.

Recommendation 5: Cooperation Is Key
Wherever possible, cooperation with local stakeholders, and engaged persons is recommended. They can be found in neighborhood management, social work, libraries, associations, schools, educational initiatives, and self-help groups—or in committed members of the target group itself (Lloyd et al., 2012, p. 55; Marschalek and Schrammel, 2017, p. 34; Smithsonian Institution Office of Policy Analysis, 2001, p. vi–vii).
They know the situation and the needs of the people and thus make insights and approaches possible in the first place. They can advise on the relevance of topics, working approaches, and avoidable pitfalls. Often, they are trusted persons for the community and their word carries weight in the group. This turns them into door openers building up trust, which is often a prerequisite to successful communication projects (Marschalek and Schrammel, 2017, p. 27, 35). This is also true for communicating via mass media. Another possibility is that researchers with connections to the target group play an important role in a communication project. (…)      
Furthermore, in the existing guidelines as well as in our work it turned out that potential cooperation partners have only very limited time and financial resources (Marschalek and Schrammel, 2017, p. 34). They understandably want to focus these few resources on their respective core tasks. Even if the requested contribution is “only” limited to counseling and facilitating access to the community, it is another project that requires at least some coordination and attention. It is therefore crucial that potential cooperation partners are not simply regarded as service providers. Their interests and their limited time should be respected and the extent to which they can benefit from cooperation in the short and long term should be taken into account.

Recommendation 6: Mind the “Openness Paradox”
The “Science for All” project started with a very open and participatory approach: (...) However, that openness made it harder for us and the participants at first, even if this sounds paradoxical. The more open the project, the more prior knowledge, and initiative the participants need. The joint development of the topic and format was difficult to communicate, whether to potential project partners or the groups themselves. This challenge is not only based on specific knowledge gaps, but maybe even more so on differences in cultural and science capital. Participatory formats are much more inviting and accessible for communities that have experienced self-efficacy in shaping their careers and being part of a political and scientific discourse.
With a concrete institution as a sender (such as a university or an association in a district), a goal defined by it (e.g., a topic-specific educational mission) and a topic set by it (which derives from the work of the institution, e.g., basic scientific education), the prerequisites would probably be more favorable.
This is true for interaction, too. Interaction is often praised as a tool for making science communication more attractive (e.g., The Science Museum, 2016; Sievert and Purav, 2018), but interaction also requires an understanding of how it works and how to make the best use of it on the participants' side. This makes it less accessible especially for people with low literacy (Dawson, 2019, p. 114).
However, this does not mean the first advice—listening—is superfluous. You should still be open to the needs and interests of the people and adapt accordingly, but at the same time not ask too much of them.

Recommendation 7: Implement Long-Term Activities
Project financing is often limited to one-off activities and pilot projects, especially driven by an increasing short-term and only project-based funding in science instead of a basic financing of independent scientific institutions. If these projects are designed accordingly, these can indeed develop an experiential character, attract attention, and create a first approach. However, if it stops there, the effect will quickly fizzle out and lead to frustration among the groups addressed. This even applies if local intermediaries support the project. It might still not be perceived as an authentic initiative from within the community and with a long-term perspective and lasting impact. In the end, it can be very time-consuming to build the trust between the science institutions and communicators and the people addressed, which is often a prerequisite to reach them (see recommendation 5).
While this recommendation would apply to many other forms of communication, it is particularly relevant for engaging underserved communities with a history of being marginalized. Dawson (2019, p. 92) shows how one-time activities—even when they are well-intentioned—can backfire by describing that one of her interview partners “argued everyday science learning activities tailored to her community during Black history month were tokenistic, angrily stating, ‘we’re not invited the rest of the year!”
It would be ideal if an institution or initiative with a concrete local reference (such as the district or the environment of a university) or group-specific contacts (self-help group, community association) started a project, instead of an intervention from the outside. This should also have at least a medium-term perspective so that the start-up phase necessary in all projects can then be used to generate further interest and participation.