How to finish a thesis
September 22, 2019 10:57 AM   Subscribe

I have been working on my master's thesis for a year, and I am stuck on how to complete it. Is there a way I can complete a thesis with only 70 responses to a survey, or does anyone know a way of targeting a specific population (of high school librarians) to gain more responses?

I am completing a masters in sports management, and my thesis title is How can high school librarians better prepare student-athletes for the rigors of higher education?.

I created a survey for high school librarians, but I am having trouble getting enough people to respond to create statistically valid data. I live in Western New York and my original plan for my thesis was to survey the high school librarians in my geographic area. After I sent my survey to the network I have access to, my advisor informed me I need to get close to 100 responses to have usable data. Over the summer, I sent the survey to over 700 email addresses I found in the New York Library Association School Library section, and still only have 70 responses.

The biggest problem I'm having, aside from lack of responses, is that in reviewing the responses I have, I'm not sure quantitative analysis is the right way to study my topic. If anyone has advice on how to finish the thesis process, I would really appreciate it. I only have one RL friend who has done this level of research, and even he thinks my school's thesis process is excessive. I'm not planning on going on to complete doctoral research, I just want to get a job in a college athletics department.

Thanks for any advice you might have. I've been a longtime Metafilter reader, but this is my first question.
posted by DEiBnL13 to Education (14 answers total)
 
Have you followed up with the people you sent it out to? I definitely need reminders, and I’ve also been the person sending out surveys (working in athletics, funnily enough) and we get a huge extra response every time we send a reminder. Especially true if you’re targeting people who work in education over the start of the academic year, when they’ll be particularly busy.

Is there a specific reason they need to be from your area? If you got a 10% return rate from your first set of requests, and you need another 30, can’t you move on to target another area and fire out 300 more?
posted by penguin pie at 11:09 AM on September 22, 2019 [1 favorite]


I used to work for an education marketing company and we bought contact information like this -- like, "high school librarians" -- from MDR. I have no idea what it costs, but relative to the kind of thing we were doing, yours is a tiny, easy request. So it might be worth just checking to see what it costs to get those email addresses? And depending on how narrow or wide you want to go, you can drill down by city, public school vs. private, and a million other ways.

And then yeah, a reminder should be sent (more than one, even!), whether you're working with your new or old data set.
posted by BlahLaLa at 11:17 AM on September 22, 2019


Yes, re-send the survey. School librarians in my area don’t work over the summer.
posted by stellaluna at 11:27 AM on September 22, 2019


Looks like you've already reached out to NYLA, but you might consider reaching out to your state library, too. Wondering if there are folks who might send on your behalf, warming up the request a bit.

I'm also wondering if your survey is onerous to fill out. How many questions/much time does it take? There could also be an issue with sending in the summer when libs have that time off (of course the opposite could be true as well, and now they're too busy).

In addition to following up with the folks who haven't filled it out, is there a way to do a little snowball sampling and ask the folks that did to reach out to their networks?

It's impossible to advise on whether or not qual/quant approach is the best without knowing more about your survey and your research questions. Tell us more.
posted by 10ch at 11:31 AM on September 22, 2019


Why don't you send it out on the AASL list serv? I see these things on there all the time.
posted by aetg at 12:00 PM on September 22, 2019 [1 favorite]


10% engagement based on sample size is actually a pretty good return for surveys, but sending out reminders may well be worth a punt. Is it a statistical issue that suggests 100 responses is needed for legitimacy? I've seen plenty of things published with less returns and that sounds questionable if 700 is the basic set to draw on.

The other alternative might be to think about doing some qualitative work on the subject. Work out some questions and contact some relevant librarians to speak to in person. You already have a list so that's a starting point. It would be a lot of work to add this method though. Potentially lots and lots of transcription.
posted by biffa at 12:03 PM on September 22, 2019


[I]n reviewing the responses I have, I'm not sure quantitative analysis is the right way to study my topic. If anyone has advice on how to finish the thesis process, I would really appreciate it.

Talk to your advisor about this. Giving you advice on how to finish the thesis process and -- especially! -- what kind of analysis is appropriate considering the nature of the data is your advisor's job.
posted by heatherlogan at 1:17 PM on September 22, 2019 [5 favorites]


Your data analysis plan should have been in place before the survey instrument was created and data collection even began. What are your research questions/hypotheses? What sort of data is your survey collecting?
posted by Young Kullervo at 1:31 PM on September 22, 2019 [3 favorites]


Yes, definitely talk with your advisor. Ultimately, the answer to this question is going to be "whatever your committee will approve as enough work to be finished".

I agree with the others that resending the survey out is the best first step. There were some great suggestions above on other places to look for survey responses. Reaching out to smaller groups/snowball might be helpful in that people might feel more invested in responding if they have some kind of emotional investment. Ideally, you'll get enough responses to analyze the data with your initial plan.

Another option, if your advisor/committee approves, would be to look at expanding your research into a mixed method approach. For example, if this topic isn't currently on the radar of high school librarians, could you talk to a small sample to find out why? Have they considered it but they have other priorities? Has it never come up specifically or in this way? Would their work on this perhaps end up grouped under other terminology (such that if you had that vocabulary to use you'd have more success in getting responses even)? But that said, biffa is spot on in saying it'd be a lot of work to add a qualitative element. Transcribing is so, so much work.

Or there's the option of expanding in another direction, and trying to get some information from people who are impacted by the interventions (college advisors of student athletes, or college athletes themselves) to see if they have any feedback to compare with current practices. But that'd be a such a significant change, it'd be a big project in and of itself.
posted by past unusual at 1:45 PM on September 22, 2019


There are some great suggestions up-thread! As others have pointed out, reminders are a must (either 1 or 2) as research shows that they increase survey response rates.

As someone who does a lot of qualitative research, I would advise you against trying to pursue this avenue last minute for your MA. Doing good qualitative research is a LOT of work and can be extremely time-consuming (plus it requires learning/training just like quant work, if you do not already have it). I would suggest employing some of the suggestions above to try and increase your response rate instead.

I'm assuming that you will be trying to do some statistical comparisons of some sort to answer hypotheses? Where does the "almost 100 responses" come from, a power calculation? Keep in mind also that plenty of master's students end up completing a thesis without hitting targets for sample size. At the end of the day, you do what you can and discuss the limitations of your work. Some research never really works out for different practical reasons. But it does need to satisfy the requirements of your master's program. Definitely do discuss this issue with not only your advisor but also your committee members. That's what they are there for, and even if they don't have any concrete suggestions that end up working, they will at least be aware of the challenges you're facing and how you're trying to address them.

Best of luck!
posted by DTMFA at 2:08 PM on September 22, 2019 [1 favorite]


I see surveys for grad work on the various librarian Facebook groups I’m on. Maybe try there?
posted by lyssabee at 3:29 PM on September 22, 2019


Thank you for all your responses! I will definitely try resending the survey to my local school library network this school year. I initially sent the survey out in June before school was out, but librarians might have been busy with end of the year tasks then. I was hesitant to send it out again because I don't want to abuse the listserv.

To answer some of the follow-up questions, the survey takes about 10 minutes in Google forms. I designed it to be relatively brief because I didn't want it to be cumbersome to people who already have a busy schedule. The questions basically ask about how librarians spend their time during the day. How often do they teach classes, how much time do they interact with students, do they offer support services to the general student population, do they work with the athletics department at all. I am targeting librarians in New York because they are an easy population for me to reach out to and we all follow the same broad New York State library curriculum. I think my research would be interesting if it included different geographic areas, especially areas where sports play a much bigger role in schools, but at this point that might be beyond the scope of my project.

I have been in communication with my thesis advisor, but she has not been entirely helpful throughout this process. It's her first time advising a master's thesis, so I understand she's learning the process too, but we need to wrap it up. She told me I should shoot for 100 responses to be statistically valid, but 75 could work.

Again, thank you for the responses. I will definitely try some of your suggestions. Sorry for the long follow-up
posted by DEiBnL13 at 6:44 PM on September 22, 2019


Incentives can be helpful to increasing survey response rates. And rather than giving an incentive to everyone (which can be expensive), people seem to go for being entered into a lottery to win some cool, unusual thing. That way, you just have to buy one thing, and tell people they need to include their email address in the survey if they want to be entered into the lottery.

Last survey I fielded we had a lottery to win an ipod, back when that was a cool thing. Because your subjects are librarians, maybe a cool reference book or unusual atlas might be a desirable incentive?

A good trick for surveys is to have your first question be interesting or slightly provocative in some way. Like don't have the first question be a demographic one!
If you can get people to answer the first question, chances are they'll stick it out to the end (unless it's too long!)

Also, having your survey be endorsed by some prestigious institution or person can get people interested in responding.
posted by jasper411 at 9:20 PM on September 22, 2019


First: about reminders; I agree that at least one reminder is ok (and totally what always happens when I get a questionnaire, because the first time around, one always thinks 'perhaps someone else will do this for me.') A little depending on the number of people involved, one also could spend a morning calling people up, the rule-of-thumb being that if you want something from people, you have to try to get attuned to their mode of communication (which involves trying various ways), instead of believing that yours 'will do.'

Then, the thing about qualitative versus quantitative. Please note that your title (your main research question is in the title, right?), as it is phrased now, is mainly addressing a qualitative problem (highlighted by the word "better"). Now, IF the data that you've accumulated so far shows CLEARLY that you're ultimately going to be getting a more relevant kind of information for answering your question by analyzing the qualitative outcomes (and, obviously, provided that your questionnaire included the possibility to add individual answers or viewpoints), 70 responses seems way too many already.

In-depth qualitative analysis means that you are "allowed" to look at fewer answers, but more carefully. You could for example [this is purely theoretically spoken since I haven't seen your material] sort everything you've got at this point into relevant categories, according to the statistical outcomes, and then pick two or so "typical" answers for each outcome and a handful of wild-card or entirely different ones, and finally give these selected answers an in-depth analysis, geared towards answering your research questions, testing your hypotheses, etc. So your statistical knowledge would allow you to sort through what you've got, whereas the ultimately selected answers provide you with qualitative results.

If you want to stay purely quantitative instead, and you are not able to get hold of more answers, you need to adjust your section that deals with your methodology, reliability-of-the-study and/or validity-of-outcomes. Putting all parts of a thesis in place is an interactive process; you must be prepared to adjust the parts that already are planned and canned as circumstances change. Not getting "enough" answers is such a change of circumstances. What does not change is the timeframe that you're supposed to deliver the work. So, to solve this conundrum, you talk about it: 'within the given timeframe, and in view of the lower-than-expected response frequency, the validity of my study will be influenced thisly and thusly. I however am confident that the results are showing this and that clearly nevertheless. Further study would be required regarding that-and-that.'
posted by Namlit at 4:14 AM on September 23, 2019


« Older Going to a work convention and trying to avoid...   |   Topic/theme/title for an English Literature BA Newer »

You are not logged in, either login or create an account to post comments