Sometimes we need to dig a bit deeper to get enough participants, or to get a more representative sample of our users. (See Dealing with selection bias later in this chapter.)

In either case, we can ask ourselves in any of these other methods would work for our particular study:

• Amazon’s Mechanical Turk
  This service lets us offer micro-payments to people to do online tasks (in this case, our tree test). But because we’re tapping into a global audience that is keen on earning a series of tiny fees quickly, we need to be extra careful about getting garbage data.
  
  
• Trade associations and customer groups
  Our customers may already have their own groups (online or not), or they may belong to a trade association (e.g. a farming collective or a plumber’s union). We should consider asking permission to contact the group’s members for our study, or ask for our invitation to be added to their next newsletter.
  
  
• Universities and other post-secondary institutions
  Students are often willing to participate in a study, both for the experience and for the chance of reward. Ask permission to post a online ad or paper bulletin.
  
  
• Targeted publications and forums
  There are message boards and online forums for every conceivable profession and interest. We can post on those that are frequented by the type of user we’re looking for, but we should first find out if we need permission first to post our request.
  
  
• Colleagues, friends, and family
  Depending on the type of user we’re looking for, we may be able to “shoulder-tap” people we know to participate. If we’re looking for graphic designers, for example, we can start with a designer we know and ask if they can pass the invitation along to other designers.
  
  
• SMS (TXT)
  If our organization has a list of mobile phone numbers for customers, we can text our study request to these people, using the same short format we may have used for Twitter invitations (see Using social media earlier in this chapter). If they have a smartphone, they can activate the link directly from the TXT message.
  We rarely use this channel, however, because (a) it usually costs money to send these messages, (b) the limited length of a plain-text message is harder to work with, and (c) recipients are more likely to consider this to be spam than if we sent the same request as an email.

Next: Dealing with selection bias