Task 2

My post is a lot of citations but I felt they all directly answered the questions.  I think I will produce more project-focused discussion in task 3.  But again, I really feel that these citations nail the questions down and also tie directly to my project (without the exact description of how, to avoid excessive post length).

1. What are the sources of bias related to sampling?
   1. “Snowball sampling relies on referrals from initial respondents to generate additional respondents.  While this technique can dramatically lower search costs, it comes at the expense of introducing bias because the technique itself substantially increases the likelihood that the sample will not be representative of the population…  Convenience sampling is generally assumed to have a higher likelihood of generating a biased sample” (Fricker, pg. 200).
   2. “Frame coverage bias occurs when the sampling frame misses some important part of the population.  For example, an email survey using a list of email addresses will miss those without an email address” (Fricker, pg. 200).
   3. Selection bias is an error in how the individual or units are chosen to participate in the survey.  It can occur, for example, if survey participation depends on the respondents having access to particular equipment, such as Internet-based surveys that miss those without Internet access” (Fricker, pg. 200-201).
   4. “Size bias occurs when some units have a greater chance of being selected than others.  For example, in a systematic sample of website visitors, frequent site visitors are more likely to get selected into the sample than those that do not.  In a similar vein, when selecting from a frame consisting of email addresses, individuals with multiple email addresses would have a higher chance of being selected into a sample” (Fricker, pg. 201).
   5. “Non-response bias occurs if those who refuse to answer the survey are somehow systematically different from those who do answer it” (Fricker, pg. 201).
2. What are two possible sampling methods for your final project?
   1. “Cluster sampling is applicable when the natural sampling unit is a group or cluster of individual units.  For example, in surveys of Internet users it is sometimes useful or convenient to first sample by discussion groups or Internet domains, and then to sample individual users within the groups or domains” (Fricker, pg. 199).  My research is taking place at several different grouping levels.  I can effectively cluster any of the following:
      1. Institution
      2. Athletic Department
      3. Athletic Team
      4. Coach
      5. Player
   2. Callegaro’s step-by-step recruitment procedure (pg. 1012); Gallup does this.  It recruits with RDD then eligible families are sent a membership packet (Callegaro, pg. 1015).  I could use this method complementary to cluster sampling.
      1. Contact and recruit the potential members via telephone (random-digit dial (RDD)), face-to-face, mail, or email.  These are pre-recruited probability-based Web panels (Callegaro, pg. 1009).
      2. Administer a “profile” or “welcome” survey in order to collect basic demographic information on the potential panelists.
      3. Determine eligible and non-eligible cases.
      4. The eligible cases are kept and put in three categories:
         1. Refusals and break-offs,
         2. Non-contacts, and
         3. Other non-interviews
   3. [I don’t think I’ll use this one.  Voluntary sampling appears to be more for marketing purposes whereas my research is targeting populations that are not already seeking to be anywhere near me or any researcher for that matter.  This is for my notes.]  Volunteer opt-in panels (Callegaro, pg. 2009) can be single or double opt-in (Callegaro, pg. 1015).  Double opt-in would involve e.g. a follow-up email confirmation with a link.  Single opt-in is the link given on first contact, e.g. the first email.  Opt-in streams come from:
      1. Word of mouth
      2. Advertising
      3. Pop-up window
3. How would you calculate the response rate for a sampling method that you may use?
   1. “IMRO (2006) gives a definition of response rate and completion rate. Response rate is ‘based on the people who have accepted the invitation to the survey and started to complete the survey. Even if they are disquali?ed during screening, the attempt quali?es as a response’ (IMRO, 2006, p. 13). Completion rate ‘is calculated as the proportion of those who have started, quali?ed, and then completed the survey’ (IMRO, 2006, p. 13).
   2. To reduce coverage error, I will “obtain as complete a sampling frame as possible (or employ a frameless sampling strategy in which most or all of the target population has a positive chance of being sampled)” (Fricker, pg. 198).

I also found these two statements intriguing:

“All online panels are generally very dynamic with members joining and leaving” (Callegaro, pg. 1016).  I would hope that leaving is not too common!

“A client may not want sampled members to have completed a recent survey on a similar topic or on a topic that might in?uence the survey outcome” (Callegaro, pg. 1017).  We have discussed influence in previous modules.  I mentioned once that images might distract the respondent and make him forget his previous answer.  Could it be that it is useful to “make the respondent forget” a prior survey altogether?  This might eliminate or reduce some lingering influence.

Interactive Marketing Research Organization. 2006. “IMRO Guidelines for Best Practices in Online Sample and Panel Management.” Available at http://www.imro.org/pdf/IMRO_Guidelines_for_Best_Practices_in_Online_Sample_and_Panel_Management.pdf (accessed September 2008).