ASK A QUESTION

In this week’s interview, the research question I chose is “information behavior related to buying your laptop”. Accordingly, non-structured interview question is “How would you decide to buy your laptop?”; semi-structured questions are: 1) What information do you consider when you decide to buy your laptop? 2) Where do you get those information? 3) Are there any other information related behaviors you’d like to talk about related to your laptop decision?; critical incident interview question is: Could you describe your last experience to buy your laptop?

In order to talk about later information searching or inquiring, most of my interviewees spent a lot of time describing their criteria and preference in laptop decision, which leaves little time for information behavior questions. Maybe because people would become talkative when it relates to their own preferences? It seems hard to drag them back to “not that interested” information related questions.

I found critical incident question is much easier for me. Because the steps of the incident can be tracked, conversations can be led to the way I want. Not sure if anyone else have the same feeling:)

1. Study questions: How users of social Q & A evaluate the quality of answers? Are there any criteria they prefer in identifying answer quality?

2.Unit of Analysis: Users of  Stack Overflow (a social Q&A site for programmers: http://stackoverflow.com/)

3.Data Collection: Data collection consist of content analysis of the comments of high-quality answers (upvote score >100), statistical analysis about the relationship between upvote and different features presented in the answers, and interviews with users.

4. Data Analysis:

a) All of the interviews will be audiorecorded, transcribed, and coded using NVivo 10. Two researchers independently code all of the interviews to identify major themes for each research question. After comparing, discussing, and resolving any differences in their coding, the researchers create a codebook and use it to recode all of the data. 

b) We will filter out unqualified comments and then identify specific accounts of “selection criteria” in the qualified comments using an inductive content analysis.  Two researchers participate in the analysis. The initial coding scheme emerge from interview findings and previous research.

c) We will divide answer feature set into 8 groups: Structure, Length, Style, Relevance, Review History, User, User Graph, and Readability. We then conduct two series of experiments. First, we represent our question-answer pairs using only the features of each group in isolation, in order to determine the individual impact of the group. Following, we represented the Q&A pairs using all the features, leaving out one group at a time. By this way, we can verify how each group is able to contribute to the results, independently from the other groups.

5. Alternative Design Possibilities: I’m very interested in using survey to examine what selection criteria users prefer in identifying high-quality answers regrading programming, because I assume the criteria they use will be different based on individual characteristics (e.g. how long they use that site, how active they are on that site, their own expertise, etc.). Survey opens up the possibilities to approach a relatively large size of studying subjects directly (comments  and statistical models can only reflect their opinions from a data perspective). However, this site does not offer in-site message function and only part of the users provide their contact information (email, etc.)on the profile pages, so sampling will be a problem if I decide to use a survey.

 

Ask a question

My hypothetical study will attempt to examine how oyster harvesters in Apalachicola Bay are handling the declining oyster harvests in the Bay. Apalachicola Bay has seen its harvests dwindle in size in recent years (Sommer, 2013) due to water use issues facing the Apalachicola-Chatahoochee-Flint river system.   These oyster harvests are a prime source of income for approximately 2,700-4,000 of Franklin County, Florida’s 11,000 residents for some portion of the year (Sommer). The “tri-state water wars” (Ritchie, 2014, para. 1) have largely been precipitated by growing populations in Georgia and Alabama, mainly in the Atlanta, Georgia area (START, 2014) as well as increased agricultural water use in Georgia and Alabama (Havens et al., 2013). My research question, then, is: How are Apalachicola Bay oyster harvesters negotiating their changing identities in the face of unprecedented economic and environmental issues.

Pick a case

The particular case I will study is oyster harvesters in Apalachicola Bay who have reported through a preliminary survey that they have participated in the oyster industry for at least one year and derive at least 25% of their yearly income from the industry. These requirements are slightly arbitrary, but having a least one year’s experience in the industry and losing 25% or more of a person’s yearly income will likely have a large effect on the life of the individual and that person’s family. The purpose of these minimum requirements for inclusion is to allow for greater insight into the changing circumstances that oyster harvesters are facing. I will attempt to interview as many oyster harvesters as possible in order to reach a saturation point of experiences.

Describe the type of case

I will use the holistic, single-case design for this study. I will use this approach because no logical subunit of analysis can readily be identified at the outset of the research. Additionally, the oyster situation in Apalachicola Bay is such a unique phenomenon that it will likely provide enough information to justify documenting and analyzing. As mentioned above, I will choose participants based on their answers to a previously distributed survey. Because I’m choosing a single-case approach, I will need to be sure that I minimize the potential for misrepresentation and that the outcomes are representative of the population under investigation. This will be done through the preliminary survey and formative research. I will use in-depth interviews to reach a saturation point. I will attempt to balance the interviews between a questioning and conversational nature.

Type of data and analysis

I will gather data that attempt to answer my overarching research question. More specifically, I will ask about specific methods that oyster harvesters are using to relieve some of the financial burdens of lost income, how their families are coping with their changing identities, how oyster harvesters are adapting their skill set to other industries in the region, and how oyster harvesters plan to help mitigate future ecological issues of a similar nature while still being able to make a living off of the bay. I will analyze my data using the constant comparative method as outlined by Corbin and Strauss (1998) to examine, compare, synthesize, and categorize the data I’ve gathered.

How could the findings differ from what they could be if I did a study on the same question but with more participants/sites/data and fewer data types?

I think the biggest change in data would occur if I conducted the study over a series of sites in Florida and Alabama. The change would be a result of the different options oyster harvesters would have if they decided to leave the industry altogether. In Alabama, as opposed to Florida, harvesters may have more local agriculture options for work, which may lend better to the skill set that they have already developed from their time in the Bay. I still think, though, that no matter how many individuals participated in the study, a saturation point of identity negotiation techniques and experiences would occur rather quickly. This would be due to the limited number of options an individual has in the region unless he or she is willing to completely uproot his or her family and move to an entirely new location.

Research Question

What are the information behaviors of individuals during a crisis situation?

Pick a Case / Unit of Analysis

This study is about the information behaviors of occupants in a house that has been damaged by a tree during a severe thunderstorm in North Florida.

Case Description

This holistic single-case design case study’s goal is to test the information behavior in crisis model with a small localized crisis situation. The information behavior in crisis model integrates information resources, information behavior and needs of the storm-affected communities within the temporal stages of a natural disaster (Lopatovska & Smiley, 2014). An autoethnographic study was conducted during Hurricane Sandy informed the authors on the building of the model and its accompanying framework. This study of a North Florida household will help determine whether the model and its framework can be branched out or extended into small localized crisis situations and find possible avenues for future research in building a theory for crisis informatics. An important consideration of adopting a holistic design is the potential to shift from the original research question while conducting the study (Yin, 2003, p. 45). The adoption of an existing model may provide insurance against drift of focus.

Types of data to collect

There are three types of data to be collected for this study. The collection of real-time data in the form of a participant observer diary. A comparison data set will be provided by interviews with individuals that also experienced the event. Finally, data from external sources such as details about the environment, researcher, data collection, and analysis steps will provide a robust account of the case.

Types of analysis to complete

Content analysis will be completed by using the information behavior in crisis model and framework to code and interpret the data. The coding themes will mirror the themes used by Lopatovska and Smiley (2014) in order to provide consistency. The coding themes are information need, information behavior, and information source.

Alternative Design Possibilities

The findings might differ if this study used the same question but contained more participants, sites, and data sources. A factor to consider is that multiple-case designs are often preferred over single-case because the more cases collected during a study will increase the chance of success (Yin, 2003, p. 53). Also, the more data that is collected the greater the possibility to correctly identify themes. It is possible that the participant observer might have some idiosyncratic information behaviors that are not common among a more diverse population. The inclusion of more geographical regions will help provide a more general understanding of the information resources needed and are available during a natural disaster. It is possible that weather conditions in California might influence information behaviors differently than the ones identified in North Florida.

References

Lopatovska, I., & Smiley, B. (2014). Proposed model of information behaviour in crisis: The case of hurricane sandy. Information Research-An International Electronic Journal, 19(1).

Yin, R. K. (2003). Designing case studies. In Case study research: Design and methods (3rd ed., pp. 19-56). Thousand Oaks, CA: Sage.

Case Study:  LGBTQ collections in rural public libraries in the American South

Questions:

1. How comprehensive and relevant are the LGBTQ collections in rural public libraries in the American South?

2.  How comprehensive and relevant do librarians in rural public libraries in the American South believe their LGBTQ collections are?

The case study described here would be a holistic multiple case study to allow for comparison not only between LGBTQ collections in comparable libraries but also the beliefs about the LGBTQ collections by the librarians responsible for building and maintaining these collections.  A single case study would be of interest but would ultimately not provide any measure of generalizability; while a multiple case study might not either, it would be more likely to, depending on the outcomes.  The units of analysis in this study would be the individual libraries.

There would be a large amount of basic catalog searching involved in answering research question 1.  Each library’s catalog would need to be searched by keyword using a specified list of terms to determine the extent of the LGBTQ materials in the collection.  After this list has been used, a title search based on popular and award-winning LGBTQ literature would also be used.  These searches will also reveal whether the library’s cataloging practices are thorough and accurate.  After determining what sort of LGBTQ collection is present, librarians at each library would be interviewed about the LGBTQ materials available (or not) in their library.

Were this study to be redesigned to include larger numbers of libraries, it would certainly paint a more comprehensive picture of this issue in this particular region.  One could certainly do a study in which the collections were analyzed for range and type of LGBTQ materials available, but to only seek librarian opinions about the state of their LGBTQ collections would be nearly useless without the collection data with which to compare them.

Case Study Design: Implementation of a Schoolwide Reading Program

Study Question: How does a school go about implementing a schoolwide reading program at the junior high school level?  What role does each employee play, and how do they feel about those roles?  What works and doesn’t work during the process?

Study Propositions: None.  Exploratory Case.

Unit of Analysis: a suburban junior high school serving grades 7-8

Data Collection: Data collection would consist of examining internal documents regarding the reading program, observation of the reading program in action, and interviews with the employees of the school (including both administration and instructional personnel).

Data Analysis:  Documents would be analyzed for content, and field notes/interview transcripts would be analyzed and coded, looking for emerging themes.

Alternative Methods:  If I were to conduct this study as, say, a survey, I would be able to gather data from a larger group of people.  I could distribute the survey electronically across the U.S. and even internationally, and get a broader view of the issue.  However, the data gathered would not be as rich as what I could get from a case study.  The case study is more narrow in scope, but would be more in depth, and possibly reveal themes that would not emerge in a survey.  Further, using multiple methods in the case study would allow for triangulation of data that a simple survey could not provide.  As you know, I plan to use a case study research design for my dissertation.  I chose the case study precisely because of it’s flexibility of methods and ability to gather rich data.

ASK A QUESTION

My hypothetical study involves college freshmen and their perceptions of information literacy. Though largely adroit in terms of communication technologies, this multifaceted group may lack information literacy proficiency (Agosto & Hughes-Hassell, 2005; Gross & Latha, 2013; Jones, 2007). They can be described as having inflated perceptions of their ability to locate, synthesize and critically assess information. College students’ “self-taught” media skills often do not match the level at which they scrutinize content (Agosto & Hughes-Hassell, 2005; Gross & Latham, 2013; Jones, 2007). Furthermore, it can be argued that college freshmen view the academic library as a place to propel them in their self-guided pursuits as opposed to a resource by which they can receive instruction or training with—perish the thought—a librarian as the facilitator. In this vein, competency theory (Kruger, J., & Dunning, D., 1999), which posits that individuals with some domain knowledge are often unable to recognize their deficiencies and therefore tend to overestimate their skill level, is used as the theoretical framework herein.

My research question are: How do FSU freshmen rate their information-literacy skills? How do students’ perceptions compare to their abilities?

PICK A CASE

The particular case that I will study is FSU freshman students who reported their information-literacy skills to be moderate to high. I will investigate the logic and information by which the students assessed themselves. The purpose is to understand how students perceive information literacy; their experience or exposure (if any) to K-12 information literacy instruction; and their current information technology proficiency. I will not set a limit on the number of students that I will interview, which leads me to the next point.

DESCRIBE THE TYPE OF CASE

I have chosen the holistic multiple-case replication design. Thus, I will replicate the “experiment” or inquiry by continuing to study students until I have reached a point of saturation. I will choose participants, or cases, based on student demographic information. My aim is to obtain a group that is representative of the population parameter. In terms of technique, I will utilize the in-depth interview and will have a questionnaire. Since I am aiming for “the replication of an experiment” (Yin, p. 52), the case study research must be standardized or formal. Thus, based on our previous interview readings, I must strike a balance between conversation and interrogation. Though a multiple-case design requires more resources and time, my research question and desired outcome calls for a comprehensive description; the multiple-case design is the way to derive exhaustive information.

TYPE OF DATA AND ANALYSIS

When it comes to data, I must first ask assessment questions in order to see if students self-perceived moderate-to-high information skills are accurate. Students should be able to briefly employ information literacy skills. Afterward, I will probe to distinguish whether the participants earnestly understand the definition and concepts of information literacy. It is also important for me to distinguish the students K-12 information literacy education. Regarding the analysis technique, I will use the constant comparative method in order to synthesize, examine, compare, conceptualize and categorize data.

HOW COULD THE FINDINGS DIFFER FROM WHAT THEY WOULD BE IF I DID A STUDY ON THE SAME QUESTION BUT WITH MORE PARTICIPANTS/SITES/DATA AND FEWER DATA TYPES?

Another way to approach this same study is via a survey study. However, survey research is often superficial and lends itself to the very phenomenon that I intend to study: college freshman’s inflated perceptions of their information literacy skills. A trajectory involving multiple case study design utilizing in-depth interviews and constant comparative analysis will allow for deep exploration and skills assessment. As Yin mentioned, “the evidence from multiple cases is often considered more compelling, and the overall study is therefore regarded as being more robust.” (p.41)

This week’s exercise was eye-opening in a couple of ways. First, I have discovered that I am not a good interviewer. I failed to have enough probes in my critical incident interview to keep my subject talking for even the five minutes required for the exercise. The topic was something I am incredibly interested in and could talk about for hours on end, but I did not account for the fact that any subjects I would have might not be as engrossed as I am. This is an excellent lesson to learn. As researchers, we are, of course, generally very involved in the topics we are studying and can be oblivious to other people’s indifference or lukewarm attitude toward the subject.

Second, as the interviewee, I became aware of the difficulty of answering pretty random questions on the fly. Hopefully, I can adjust for my own shortcomings in the area of interviews, because the plan is to have interviews in my dissertation and some other research.

Each of the three interview instruments (i.e. unstructured, critical incident, and semi-structure) have their own strengths. An unstructured interview allows for the discovery of topics that the interviewer might not conceive of on their own. For example I asked about online video viewing habits but had not thought about the primary language of the video shaping the choice of website. I found that this technique was the easiest for me to keep the interviewee talking for the allotted time. A factor that allows this to occur is that the specifics of the topic are controlled by the interviewee. This makes sense since the interviewee is most likely not to choose a direction that is of little interest to themselves.

I found the example driven element of the critical incident interview to be helpful in discovering details. This instrument provides a storytelling aspect that brings the topic to a more concrete example that encourages talking about behaviors in a situation rather than a more conceptual approach on how someone looks for videos. For example, I found that someone watched the same video 4 times in a row. This aspect of watching a video multiple consecutive times did not occur to me even though I engage in this very behavior myself.

The semi-structure interview was the most difficult. My troubles started with the formulating of the questions. I found that I am most comfortable with quantitative question forms. These are the type of questions that illicit yes/no, a category, or some point on a Likert scale. Thankfully I was provided an example, this allowed me to imitate the form and insert my chosen topic. Even with the more open-ended form I found it hard to keep the interviewee talking. My question writing is probably not to blame since it was the last interview round. My interviewee and I were showing signs of fatigue and starting to mentally move towards the next part of class.

In general, I found this activity gave me a taste of what is involved in interviewing that I would not necessarily understand from reading a chapter in a book. Many sources stress the need to take notes immediately after the interview. As demonstrated in class, many of us did not want to take the time for note taking and jumped into the next interview.