Qualitative Research 質性研究/定性研究 12 Qualitative Research 質性研究/定性研究 中央大學.資訊管理系 范錚強 mailto: ckfarn@mgt.ncu.edu.tw 2013.05 updated

質性研究 研究者本身作為研究的工具 (instrument) 透過其針對研究對象的訪談、觀察、文件分析等，經由資料的詮釋，獲得進一步的瞭解 中央資管：范錚強

Why Use Qualitative Research? “Most of what influences what we say and do occurs below the level of awareness. That’s why we need new techniques: to get at hidden knowledge – to get at what people don’t know they know.” Jerry Zaltman Qualitative research includes an array of interpretative techniques which seek to describe, decode, translate, and otherwise come to terms with the meaning, not the frequency, of certain more or less naturally occurring phenomena in the social world. Qualitative research can be used in both the data collection and data analysis stages of a research project. 中央資管：范錚強

Qualitative Research Data Collection Techniques Group Focus Groups Interviews Focus Groups Observation Data Collection Techniques IDIs Ethnography Case Studies Action Research Grounded Theory This slide highlights many of the qualitative techniques that are useful for data collection. 中央資管：范錚強

Behavioral Observations Qualitative Research Trace Evidence Artifacts Other Techniques Behavioral Observations Textual Analysis Debriefings This slide highlights many of the qualitative tools useful for data collection or data analysis. 中央資管：范錚強

Qualitative Research in Business Job Analysis Advertising Concept Development Productivity Enhancement New Product Development Benefits Management Retail Design Process Understanding Union Representation Market Segmentation Sales Analysis Exhibit 7-1 lists some uses of qualitative research in business. The full exhibit is provided below. 中央資管：范錚強

Artifacts/ media products Data Sources People Organizations Texts Environments Artifacts/ media products Qualitative research draws data from people and organizations. Whether the source is people or organization, we can use their behavior, texts, events and so on as data. Chapter 9 focuses on observation methods. Events and happenings 中央資管：范錚強

Distinction between Qualitative & Quantitative Theory Building Theory Testing This slide reflects information from exhibit 7-2 Quantitative research is the precise count of some behavior, knowledge, opinion or attitude. While the survey is not the only quantitative method, it is the dominant one. Quantitative research is often used for theory testing. For example, it might answer the question “Will a $1-off instant coupon or a $1.50 mail-in rebate generate more sales for Kellogg’s Special K?” It requires that the researcher maintain a distance from the research so as not to bias the results. Qualitative research is sometimes called interpretive research because it seeks to develop understanding through detailed description. It builds theory but rarely tests it. Several key distinctions exist between qualitative and quantitative research and these are elaborated on in Exhibit 7-2. The next several slides highlight these distinctions. 中央資管：范錚強

瞭解研究問題 實質問題 形式 研究問題的形式，提供了選擇適當研究策略的提示 我的研究到底做些什麼？what is my study about? 形式 Am I asking a “who,” “what,” “where,” “why,” or “how” question? 研究問題的形式，提供了選擇適當研究策略的提示 中央資管：范錚強

研究策略的選擇 研究的最重要的的一步：定義研究問題 判定研究問題的類型 一般而言： 什麼 (what)：探索性 (exploratory) 如何 (how) 、為何 (why)：驗證 (confirmatory) 中央資管：范錚強

實徵性Empirical 研究分類 探索性和質性研究有很高的相關性 依研究階段 探索性：exploratory hypothesis building （達爾文） 驗證性：confirmatory hypothesis testing 依實徵資料蒐集方法分類 質性研究 Qualitative Research 個案研究，田野調查，參與觀察，歷史探索，行動研究 定量研究（數量化研究）Quantitative Research 問卷調查，實驗室實驗，實地實驗，內容分析 探索性和質性研究有很高的相關性 中央資管：范錚強

X Y x y 探索：構念的形成 抽象化 構念化 構念層次 推論：因果關係 因果、共變關係 實質層次 測量 誤差 誤差 測量 中央資管：范錚強 實質層次

The Roots of Qualitative Research Economics Psychology Sociology Semiotics 符號學 Anthropology Qualitative research methodologies have roots in a variety of disciplines. These are named in the slide. Some believe that qualitative data are too subjective and susceptible to human error and bias in data collection and interpretation. The fact that results cannot be generalized from a qualitative study to a larger population is considered a fundamental weakness. Despite these limitations, managers are returning to these techniques as quantitative techniques fall short of providing the insights needed to make those ever-more-expensive decisions. Managers must deal with the issue of trustworthiness of qualitative data using the following techniques: Using literature searches to build probing questions, Justifying the method chosen, Using a field setting, Choosing sample participants for relevance rather than representation of target population, Using questions that will find the exception to the rule, Carefully structuring the data analysis, Comparing data across multiple sources and contexts, And conducting peer-researcher debriefing on results for added clarity, insights, and reduced bias. Communication 中央資管：范錚強

社會科學研究哲學 Positivism 實證主義 康德 (Comte) the only authentic knowledge is that which is based on sense experience and positive verification. Interpretivism or Anti- positivism/Hermeneutics 詮釋主義 韋伯 (Max Weber) ，黑格爾 (Hegel)： 辯證法 the study of the interpretation of written texts (also, verbal and nonverbal forms of communication) non-positivist research is usually qualitative, while positivist research is more quantitative. Interactionism meaning is produced through the interactions of individuals.. 中央資管：范錚強

質性研究的理念分類1 傳記研究 Biography 現像學 Phenomenology 紮根理論 Grounded Theory 研究個人，以及其陳述的經驗 現像學 Phenomenology 研究多個個人，以及他們所陳述的經驗 紮根理論 Grounded Theory 社會學：針對一個特定情境，試圖從現像中推演理論 民俗誌 Ethnography 人類學：針對一個文化的描述和詮釋 中央資管：范錚強

質性研究的理念分類2 個案研究 行動研究 歷史研究 針對一個或多個對象的描述和詮釋 組織發展：研究者參與到行動中以規劃、實施、監測行動的變化，並利用研究者的理論與經驗服務於被研究對象 歷史研究 詮釋事件之間的關連 中央資管：范錚強

質性研究的資料蒐集方法 次級資料 觀察 訪談 個人、群體 中央資管：范錚強

質性研究的資料分析方法 程序： Data collection Data reduction Data display Conclusion drawing and Verification 中央資管：范錚強

Focus of Research Qualitative Understanding Interpretation Quantitative Description Explanation This slide reflects information from exhibit 7-2 As mentioned in the previous slide, quantitative research is used to describe and explain. It can also be used to predict. However, qualitative research is focused on understanding and interpretation. 中央資管：范錚強

Researcher Involvement Qualitative High Participation-based Quantitative Limited Controlled This slide reflects information from exhibit 7-2 Researcher involvement in quantitative research should be minimal lest bias be introduced. However, in qualitative research, the researcher must have a high level of involvement to probe for understanding. In quantitative research, for instance, participants may never see or speak to a member of the research team. They may simply answer a self-administered survey. In qualitative research, participants may be interviewed by the researcher or spend several hours with the researcher. 中央資管：范錚強

Research Design Qualitative Longitudinal Multi-method Quantitative Cross-sectional or longitudinal Single method This slide reflects information from exhibit 7-2. Quantitative studies are usually single mode. In other words, they will usually rely on one data collection technique whether it be a telephone survey, email survey, or experiment. However, qualitative studies may use several methods in one study to increase the researcher’s ability to interpret and justify the results. 中央資管：范錚強

Sample Design and Size Qualitative Non-probability Purposive Small sample Quantitative Probability Large sample This slide reflects information from exhibit 7-2. Quantitative studies prefer samples greater than 200 and samples that are representative of the target population. Not all quantitative studies meet these criteria but these are desirable. Qualitative studies rely on small sample sizes – less than 25 people is common. The emphasis on selecting the sample is to include people with heterogeneous opinions, attitudes, and experiences. 中央資管：范錚強

Data Type and Preparation Qualitative Verbal or pictorial Reduced to verbal codes Quantitative Verbal descriptions Reduced to numeric codes This slide reflects information from exhibit 7-2. How qualitative and quantitative researchers would treat these ads as research would be very different. Qualitative research can also use software to conduct content analysis but words and pictures are used as codes, rather than numbers. The researcher would take the copy and images in these ads and look for themes and patterns…for example, that they all contain people, that they all contain Web URLs, that one of three is promoting a particular proprietary research service while the others are more general. Quantitative studies take verbal descriptions of consumer behavior, attitudes, and opinions and they use numbers to represent those descriptions in a database. The researcher would take the copy and images of these ads and code them with numbers. People in ads would get a 1 for male, 2 for female, 3 for indeterminate gender. Ad themes might be “1” for proprietary research service (Conceptor for Decision Analyst), “2” for institutional theme (like ‘curiosity’ for Synovate), a “3” might be assigned for general capabilities (like qualitative research services for Harris Interactive). 中央資管：范錚強

Turnaround Qualitative Shorter turnaround possible 但是，總期程長 Insight development ongoing Quantitative May be time-consuming (for a sample) 總期程短 Insight development follows data entry This slide reflects information from exhibit 7-2. Quantitative studies are traditionally time-consuming, but new methods such as web surveys are allowing for fast turnaround. The key is to recognize whether those methods are appropriate for the study at hand. Qualitative research can be faster due to the small sample sizes, but coding and analyzing hours of interviews can also be time consuming. One advantage of qualitative research is that insight development goes on throughout the study so interviews can be stopped when the appropriate answers are identified. This is not the case with quantitative studies. 中央資管：范錚強

Data Analysis Qualitative Nonquantitative; human Judgment mixed with fact Emphasis on themes Quantitative Computerized analysis Facts distinguished Emphasis on counts This slide reflects information from exhibit 7-2. Quantitative data analysis is conducted using statistical software programs such as SAS, SPSS, or Jump. The analysis focuses on the facts identified in the study. Qualitative research is not coded into numeric values. Human interpretation and judgment are critical in creating insight from the data. Content analysis...especially with the development of software like XSight...is a primary computerized analytical approach. It is far more than a count of words; such software can help reveal themes and underlying emphasis within texts. When researchers work with focus group and IDI transcripts, the content analysis software can assist the moderator in debriefing. The ability of video to be 'marked' with such software as Video Marker from FocusVision makes the analytical process better able to link interpretations to specific content from a qualitative method participant. 中央資管：范錚強

Qualitative Research and the Research Process Exhibit 7-3 Exhibit 7-3 Is reintroduced here as a means of review and connection, after going into detail on the process issues. 中央資管：范錚強

Pretasking Activities Use product in home Bring visual stimuli Create collage Keep diaries Much of qualitative research involves the deliberate preparation of the participant, called preexercises or pretasking. This step is important due to the desire to extract detail and meaning from the participant. A variety of creative and mental exercises draw participants’ understanding of their own thought processes and ideas to the surface. Some of these are listed on the slide. Pretasking is rarely used in observation studies and is considered a major source of error in quantitative studies. Draw pictures Construct a story 中央資管：范錚強

Choosing a Qualitative Method Project’s purpose Researcher characteristics Factors Schedule Types of participants Budget The researcher chooses a qualitative methodology based on the project’s purpose, its schedule including the speed with which insights are needed; its budget, the issue(s) or topic(s) being studied; the types of participants needed; and the researcher’s skill, personality, and preferences. Topics 中央資管：范錚強

NonProbability Sampling Purposive Sampling Snowball Sampling Convenience Sampling Sample sizes for qualitative research vary by technique but are generally small. A study might include just two or three focus groups or a few dozen individual depth interviews. Qualitative research involves non-probability sampling, where little attempt is made to generate a representative sample. There are several common types. Purposive sampling means that the researchers choose participants arbitrarily for their unique characteristics or their experiences, attitudes, or perceptions. Snowball sampling means that participants refer researchers to others who have characteristics, experiences, or attitudes similar to or different from their own. Convenience sampling means that researchers select any readily available individuals as participants. 中央資管：范錚強

Qualitative Sampling General sampling rule: You should keep conducting interviews until no new insights are gained. The general sampling guideline for qualitative research is to keep sampling as long as your breadth and depth of knowledge of the issue under study is expanding, and stop when you gain no new knowledge or insights. In other words, a qualitative researcher will stop sampling when he or she has reached data redundancy. 中央資管：范錚強

The Interview Question Hierarchy Exhibit 7-6 Interviewing requires a trained interviewer (often called a moderator for group interviews). The interviewer must be able to make participants feel comfortable and probe for details without upsetting the participants. The actual interviewer is usually responsible for generating the interview or discussion guide, the list of topics to be discussed, or the questions to be asked, and in what order. In building this guide, many interviewers employ a hierarchical questioning structure. This structure is shown in Exhibit 7-6. Broader questions start the interview, designed to put participants at ease and give them a sense that they have a lot to contribute, followed by increasingly more specific questions to draw out detail. 中央資管：范錚強

Interviewer Responsibilities Recommends topics and questions Controls interview Plans location and facilities Proposes criteria for drawing sample Writes screener Recruits participants Develops pretasking activities Prepares research tools Supervises transcription Helps analyze data Draws insights Writes report The interviewer is generally responsible for many tasks related to the interview. Several of these tasks are listed in the slide. 中央資管：范錚強

Elements of a Recruitment Screener Heading Screening requirements Identity information Introduction Security questions Demographic questions Behavior questions Lifestyle questions Attitudinal and knowledge questions Articulation and creative questions Offer/ Termination One of the tasks listed in the last slide was that of writing the recruitment screener. The recruitment screener is a semistructured or structured interview guide designed to assure the interviewer that the prospect will be a good participant for the planned qualitative research. Exhibit 7-7 provides the various elements necessary for a comprehensive recruitment screener. Each question is designed to reassure the researcher that the person who has the necessary information and experiences, as well as the social and language skills to relate the desired information, is invited to participate. 中央資管：范錚強

Interview Formats Unstructured Semi-structured Structured 中央資管：范錚強 In an unstructured interview, there are no specific questions or order of topics to be discussed. Each interview is customized to each participant. In a semistructured interview, there are a few standard questions but the individual is allowed to deviate based on his or her answers and thought processes. The interviewer’s role is to probe. In a structured interview, the interview guide is detailed and specifies question order, and the way questions are to be asked. These interviews permit more direct comparability of responses and maintain interviewer neutrality. Most qualitative research relies on the unstructured or semistructured interview format. The next slide highlights the differences between unstructured or semistructured and structured interviews. 中央資管：范錚強

Requirements for Unstructured Interviews Developed dialog Distinctions Probe for answers Interviewer creativity Unstructured or semi-structured interviews rely on developing a dialog between interviewer and participant. Without this dialog and comfort between the two people, the interview will not result in valuable data. Because the researcher is seeking information that the participant may not be willing to share or may not even recognize consciously, the researcher must be creative. Further, interviewer skill is necessary to extract more and a greater variety of data. Finally, interviewer experience and skill generally result in greater clarity and more elaborate answers. Interviewer skill 中央資管：范錚強

The Interview Mode Individual Group 中央資管：范錚強 The interview is the primary data collection technique for gathering data in qualitative methodologies. Interviews may vary based on the number of people involved during the interview, the level of structure, the proximity of the interviewer to the participant, and the number of interviews conducted during the research. An interview can be conducted in groups or individually. Exhibit 7-5 compares the individual and the group interview as a research methodology. Both are important in qualitative research. This exhibit is provided on the next slide. 中央資管：范錚強

IDI vs Group Individual Interview Group Interview Research Objective Explore life of individual in depth Create case histories: repeated interviews over time Test a survey Orient the researcher to a field of inquiry and the language of the field Explore a range of attitudes, opinions, and behaviors Observe a process of consensus and disagreement Topic Concerns Detailed individual experiences, choices, biographies Sensitive issues that might provoke anxiety Issues of public interest or common concern Issues where little is known or of a hypothetical nature Participants Time-pressed participants or those difficult to recruit (e.g., elite or high-status participants) Participants with sufficient language skills (e.g., those older than seven) Participants whose distinctions would inhibit participation Participants whose backgrounds are similar or not so dissimilar as to generate conflict or discomfort Participants who can articulate their ideas Participants who offer a range of positions on issues Exhibit 7-5 中央資管：范錚強

Types of Research Using IDIs 個人深入訪談 Oral histories Sequential interviewing Types In-depth interviews Life histories Cultural interviews Critical incident techniques Exhibit 7-8 An individual depth interview (IDI) is an interaction between an individual interviewer and a single participant. Individual depth interviews generally take between 20 minutes and 2 hours to complete, depending on the issues and topics of interest and the contact method used. Some techniques, such as life histories, may take as long as 5 hours. Exhibit 7-8 highlights some types of research using IDIs. Oral histories (narratives) ask participants to relate their personal experiences and feelings related to historical events or past behavior. Cultural interviews ask participants to relate his or her experiences with a culture or subculture. Life histories extract from a single participant memories and experiences from childhood to the present day regarding a product or service category, brand, or firm. In a critical incident technique, the participant describes what led up to the incident, what he or she did or did not do, and the outcome of the action. Convergent interviewing involves experts as participants in a sequential series of IDIs. Sequential interviewing approaches the participant with questions formed around an anticipated series of activities. Ethnography involves a field-setting and unstructured interview. Grounded theory uses a structured interview but adjusts each interview based on findings from those that came before. Ethnography 中央資管：范錚強

Projective Techniques MET Laddering Association Semantic Mapping Data Collection Techniques Sentence Completion Sensory sorts Cartoons Component Sorts Thematic Apperception Within interview structures, projective techniques may be used to identify hidden or suppressed meanings. Some projective techniques are named in the slide. In word or picture association, participants are asked to match images, experiences, emotions, products, services, people, and places to whatever is being studied. In sentence completion, participants are asked to complete a sentence. In cartoons or empty balloons, participants are asked to write the dialog for a cartoonlike picture. With the Thematic Apperception Test, participants are confronted with a picture and asked to describe how the person in the picture feels and thinks. In component sorts, participants are presented with flash cards containing component features and asked to create a new combination. In sensory sorts, participants are presented with scents, textures and sounds, and asked to arrange them by one or more criteria. With semantic mapping, participants are presented with a four-quadrant map where different variables anchor the two different axes; they then spatially place brands with the four quadrants. This can also be called brand mapping. With laddering (also called benefit chaining), participants are asked to link functional features to their physical and psychological benefits, both real and ideal. MET or metaphor elicitation technique uses images to encourage participants to share their innermost feelings about a topic. These techniques are time-consuming to apply and analyze. This is a good time to introduce the student to the MET interview segments on their DVD, assign a MET interview project, or otherwise use the video to talk about the interview process involved in these techniques. Imagination Exercises 中央資管：范錚強

個人深入訪談 雙向對談 評估 可中斷訪談──降低成本 可獲得深入且大量細節的研究資料，比一其他方法獲得的資料品質更佳 可獲得詮釋資料的協助 更好的控制：可以因受訪者不同，而調整用語 成本高：時間和金錢 可中斷訪談──降低成本 但，如果該樣本是關鍵性的… 中央資管：范錚強

成功的個人訪談 受訪者具有研究所需的資料 受訪者瞭解其所扮演的角色 有足夠的誘因使受訪者配合 過濾性的問題──資訊層次 訪談者的角色 確定繼續訪談受訪者 訪談者的角色 解釋 輔導 鼓勵 中央資管：范錚強

訪談技巧 (I) 增加受訪者的配合意願 開始：破冰 萬一受訪者正在忙，或是不在… 建立良好的關係 相信過程會是愉快和令人滿意的 認為回答問題是重要且值得的 能達成一些心理滿足感 開始：破冰 萬一受訪者正在忙，或是不在… 建立良好的關係 愉快 信任 中央資管：范錚強

訪談技巧(II) 蒐集資料──刺探 訪談的紀錄 選擇和訓練訪談人 簡單的表明瞭解和興趣 一個期望的表明、暫停 重複問題 重複受訪者的答案──確認 中性的問題，或評論 澄清性問題──追問 訪談的紀錄 筆記、錄音、錄影 先徵求當事人同意 選擇和訓練訪談人 假設這是個大規模的研究，有眾多訪談者 中央資管：范錚強

工具（問卷） 是的，訪談也有問卷 階段一：發展問卷的設計策略 階段二：建構問題和測量方法 階段三：草擬問卷和修改問卷 結構性的問卷 非結構性的檢查清單 階段一：發展問卷的設計策略 階段二：建構問題和測量方法 階段三：草擬問卷和修改問卷 中央資管：范錚強

問卷的用語/文字使用 問題是否以共通的用語來呈現 問卷中使用的用語是否沒有多重意義 問題是否隱含一些未被證實或扭曲的假設 避免混淆的用語 問題是否隱含一些未被證實或扭曲的假設 問題中是否有可能偏誤的用詞 該問題是否適當的個人化 不同的受訪者族群，可能需要不同的用語 若有選項，選項是否完整涵蓋所有可能性 中央資管：范錚強

開始和受訪者的過濾 讓受訪者有配合的誘因 適當揭露訪談中問題的大致方向 告知所需的時間 可能包含過濾性的問題 揭露研究單位、委託單位和研究目標 建立開始的接觸 中央資管：范錚強

終止訪談 問題： 當受訪者沒有正確的回答過濾性的問題，是否或如何終止訪談？ 當受訪者決定停止訪談時，如何終止？ 中央資管：范錚強

資料的獲取和整理 記錄整理 文字的分析 詮釋 筆記─文稿整理 錄音帶─轉錄成文字稿 模版比對：找尋特殊 pattern 內容分析 語幹分析 Gaining insight 中央資管：范錚強

Projective Techniques 需要找出隱藏的意義，無法直接經由問答得到。可採用投射方法 p.173/174 Refer to PicProfile page 169. 中央資管：范錚強

Group Interviews Dyads Triads Mini-Groups Small Groups (Focus Group) Supergroups Group interviews involve a single interviewer with more than one research participant. They vary widely in size. Mini-groups involve 2-6 people. Small groups usually include 6-10 people and are generally the most used. Supergroups include up to 20 people. The focus group is a type of small group (6-10). It is discussed further on other slides. Dyads are frequently used when the special nature of the relationship is needed to stimulate frank discussion on a sensitive topic. Group interaction is desirable but time constraints still limit extracting detail from each participant. It is also difficult to recruit, arrange, and coordinate group discussions. 中央資管：范錚強

Determining the Number of Groups Scope Number of distinct segments Desired number of ideas Desired level of detail The skilled researcher helps the sponsor determine an appropriate number of group interviews to conduct. This slide lists the considerations affecting the number of group interviews. The following guidelines apply. The broader the issue(s), the more groups needed. The larger the number of distinct market segments of interest, the more groups needed. The larger the number of desired new ideas, the more groups needed. The greater the level of detail desired, the more groups needed. The greater the influence of ethnic and geographic differences, the more groups needed. The less homogeneity, the more groups needed. The general rule is that one should keep conducting focus groups until no new insights are gained. Level of distinction Homogeneity 中央資管：范錚強

Group Interview Modes Face-to-Face Telephone Online Videoconference The focus group is a panel of people (usually 6-10 people), led by a trained moderator, who meet for 90 minutes to 2 hours. The facilitator uses group dynamics principles to focus or guide the group in an exchange of ideas, feelings, and experiences. The term focus group was first coined by R.K. Merton in his 1956 book, The Focused Interview. Focus groups can be conducted using various modes. Telephone focus groups are effective when it is difficult to recruit the desired participants, when target group members are rare, when issues are sensitive, and when one needs national representation with a few groups. Telephone focus groups are usually shorter than traditional groups and less expensive. They should not be used when participants need to handle a product that cannot be sent ahead to them, when the session must be long, or when the participants are children. Online focus groups are very effective with teens and young adults. Access and speed are strengths of this mode, but it is more difficult to gain insight from group dynamics. Videoconferencing is likely to grow as a focus group mode because it saves time and money while creating less barrier between moderator and participants than the telephone. All methods provide for transcriptions of the full interview. These are analyzed using content analysis. 中央資管：范錚強

Selecting the Data Collection Method Exhibit 8-2 中央資管：范錚強

Selecting an Observation Data Collection Approach Exhibit 8-3 illustrates the selection of an observation data collection approach. This is discussed further on the following slide. 中央資管：范錚強

觀察研究 定性研究的一環 可以和其他方式一起進行 如︰訪談、問卷、二手資料等 多為研究的探索性階段 中央資管：范錚強

觀察研究 非行為的觀察 行為的觀察 記錄分析 實質狀況分析 實質過程分析 非語言的分析 語言的分析（包含語言外的分析） 空間的分析 中央資管：范錚強

優點 有時是唯一可行的方法 在事件發生時，即時蒐集資料 掌握可能被遺漏的資料 可在整個事件發生的自然情境下蒐集資料 該情境接受觀察的干擾 如︰福爾摩斯 可在整個事件發生的自然情境下蒐集資料 該情境接受觀察的干擾 中央資管：范錚強

限制 事件發生的不可預測性 緩慢而昂貴 只能掌握到「檯面上」的動作和表面的情報 主觀的判斷和記錄 無法用以了解已經過去的事件 中央資管：范錚強

進行觀察研究 觀察研究的種類 內容的規範 觀察者的訓練 資料蒐集 (Who; What; When; How) 簡單或有系統性的觀察 觀察研究的結構 完全非結構性－－自然情境 非結構性－－實驗室 結構性－－自然情境 完全結構性－－實驗室 內容的規範 觀察的目標 觀察者的訓練 觀察者間的一致性 資料蒐集 (Who; What; When; How) 中央資管：范錚強

選擇觀察者的準則 能在很多干擾的情境下專注 能記憶很多經歷事件的細節 能在觀察情境下不引起注意 能在觀察的資料中萃取最多的insight—抽象化能力 中央資管：范錚強

Research Design Task Details Who? What? Where? (event or time) How? When? The data collection plan specifies the details of the task. Who? The plan must specify what qualifies a person to participate as a subject of the observation. It must also assign responsibilities on the research side. What? The characteristics of the observation must be set as sampling elements and units of analysis. This is achieved when event-time dimension and “act” terms are defined. In event sampling, the researcher selects certain elements, behavioral acts, or conditions to record that answer the investigative questions. In time sampling, the researcher must choose among a time-point sample, continuous real-time measurement, or a time-interval sample. For a time-point sample, recording occurs at fixed points for a specified length. Time-interval sampling records every behavior in real time but counts the behavior only once during the interval. What constitutes an “act” depends on the study. For instance, acts could include a single expressed thought, a physical movement, a facial expression, or a motor skill. When? When specifies whether the time of the study is important, and if so, what time period will be used. How? Will the data be observed directly? How will various situations be handled? How will data be recorded? Where does the act take place? 中央資管：范錚強

觀察資料的蒐集 Who What 事件抽樣 時間抽樣 When How Where 中央資管：范錚強

Content of Observation Factual Inferential Introduction/identification of salesperson and customer. Credibility of salesperson. Qualified status of customer. Time and day of week. Convenience for the customer. Welcoming attitude of the customer Product presented. Customer interest in product. Selling points presented per product. Customer acceptance of selling points of product. Number of customer objections raised per product. Customer concerns about features and benefits. Salesperson’s rebuttal of objection. Effectiveness of salesperson’s rebuttal attempts. Salesperson’s attempt to restore controls. Effectiveness of salesperson’s control attempt. Consequences for customer who prefers interaction. Length of interview. Customer’s/salesperson’s degree of enthusiasm for the interview. Environmental factors interfering with the interview. Level of distraction for the customer. Customer purchase decision. General evaluation of sale presentation skill. Exhibit 8-6 中央資管：范錚強

非干擾性的量測 反應性的線索 非干擾性的測量 物理現象的痕跡 其他方法的極佳輔助 間接測量 歷史資料搜尋 觀察分類 磨損、消耗、風化 累積（如︰垃圾學） 其他方法的極佳輔助 中央資管：范錚強

Data Collection Watching Listening Touching Smelling Reading 中央資管：范錚強 Besides collecting data visually, observation involves listening, reading, smelling, and touching. Refer back to Exhibit 8-2, which describes the conditions under which observation is appropriate. 中央資管：范錚強

Using Observation Systematic planning Properly controlled Consistently dependable Simple observation is unstructured and not standardized. Systematic observation is structured and uses standardized procedures to observe participants or objects. Observation may be the primary methodology used to answer a research question when it is systematically planned and executed, uses proper controls, and provides a consistently dependable (reliable) and accurate and authoritative (valid) account of what happened. Accurate account of events 中央資管：范錚強

Observation Classification Nonbehavioral Physical condition analysis Process or Activity analysis Record analysis Behavioral Nonverbal Linguistic Extralinguistic Spatial Observation includes the full range of monitoring behavioral and nonbehavioral activities and conditions. These can be classified as behavioral and nonbehavioral observations. Nonverbal observation is the most prevalent and refers to recording physical actions or movements of participants. These behaviors can be measured with the human eye and with several mechanical or digital devices. These devices are discussed on the following slide. Linguistic observation is the observation of human verbal behavior during conversation, presentation, or interaction. Extralinguistic observation is the recording of vocal, temporal, interaction, and verbal stylistic behaviors of human participants. It is discussed further on slide 10-12. Spatial observation is the recording of how humans physically relate to one another. Nonbehavioral observation is the observation of the effects or traces of prior actions or of nonhuman activity. Physical condition analysis is the recording of observations of current conditions resulting from prior decisions. Process (activity) analysis is observation by a time study of stages in a process, evaluated on both effectiveness and efficiency. Record analysis is the extraction of data from current or historical records. Data mining is a type of record analysis, which is discussed further in the slide show. 中央資管：范錚強

Nonbehavioral Observation Record Analysis Physical Condition Analysis Record analysis…think data mining…is a common nonbehavioral observation. What you are looking at is an accounting software report relating to late payments of customers, but such observation could as easily be looking at a defect report from operations, a record of sick days in human resources, a promotions sales-lift report, or many others that firms—large and small—generate on a regular basis. Physical condition analysis might relate to a safety audit or an analysis of inventory conditions or an analysis of food preparation areas in a restaurant. Building inspectors do physical condition analyses when they allow a builder to progress to the next phase of construction following an inspection of plumbing, wiring, etc. Process analysis started with time-motion studies in manufacturers, but today relates to any organization that tracks time related to activity or process steps and stages. Some examples include: banks (ATMs resulted from such a process analysis), merchandise retailers (e-commerce sites are constantly evaluating abandon carts), restaurants, medical practices, and architects use such nonbehavioral evaluations. Physical Process Analysis 中央資管：范錚強

Behavioral Observation “We noticed people scraping the toppings off our pizza crusts. We thought at first there was something wrong, but they said, ‘We love it, we just don’t eat the crust anymore.” Tom Santor, Donatos Pizza Donatos Pizza discovered something about consumer preferences from simple observation of human behavior. 中央資管：范錚強

Systematic Observation Standardized procedures Structured Trained observers Systematic Encoding observation information Recording schedules Systematic studies employ standardized procedures, trained observers, schedules for recording, and other devices for the observer that reflect the scientific procedures of other primary data methods. 中央資管：范錚強

Flowchart for Checklist Design Exhibit 8-5 中央資管：范錚強

Participative Observation 觀察者和觀察標的的關係 直接或是間接觀察 間接例︰單面鏡外的心理學家？ 觀察者的身分是否被受觀察者所知道 隱瞞或公開 觀察者的角色 參與者或純粹觀察 局外觀察者 現場觀察者 參與觀察者 Participative Observation 局內人觀察 中央資管：范錚強

Extralinguistic Observation Vocal Temporal Interaction Verbal Stylistic Linguistic observation is the observation of human verbal behavior during conversation, presentation, or interaction. Extralinguistic observation, the focus of this slide, is the recording of vocal, temporal, interaction, and verbal stylistic behaviors of human participants. Vocal behaviors include pitch, loudness, and timbre. Temporal behaviors include the rate of speaking, duration of utterance, and rhythm. Interaction includes the tendencies to interrupt, dominate, or inhibit. Verbal stylistic behaviors include vocabulary and pronunciation peculiarities, dialect, and characteristic expressions. Ronald Reagan was lauded for his command of extralinguistic behavior. 中央資管：范錚強

記錄 記憶 每天晚上回去整理成筆記 筆記、速記 錄音、錄影 中央資管：范錚強

Errors Introduced by Observers Halo Effect Observer Drift Inexperience can be an advantage if there is a risk that experienced observers may have preset convictions about the topic or if prior observations will influence what is perceived in a current observation. This is called the halo effect. Observers can also introduce error when fatigued, which can result in observer drift. Observer drift is error caused by decay in consistency and accuracy on recorded observations over time, affecting categorization. 中央資管：范錚強

Evaluation of Behavioral Observation Strengths Securing information that is otherwise unavailable Avoiding participant filtering/ forgetting Securing environmental context Optimizing naturalness Reducing obtrusiveness Weaknesses Enduring long periods Incurring higher expenses Having lower reliability of inferences Quantifying data Keeping large records Being limited on knowledge of cognitive processes This slide lists the strengths and limitations of observation as a data collection method. Observation is the only method available for gathering certain types of information. Another advantage is that it can capture the whole event as it occurs in its natural environment. Observation participants seem to accept an observational intrusion better than they respond to the intrusion of survey takers. Further, some observation studies are concealed A key limitation of observation is that it records what, where, who, and how, but cannot record why people behave as they do. Observation is slow and expensive. Observation may be factual or inferential. It is most reliable when it is based on fact. Exhibit 9-4, shown on the next slide, shows how we can separate the factual and inferential component of a salesperson’s presentation. 中央資管：范錚強

Combining Qualitative Methodologies Case Study Action Research Qualitative methods can be combined to glean more and better information. The case study, also called a case history, combines individual or group interviews with record analysis and observation. Researchers extract information from company brochures, annual reports, sales receipts, and newspaper and magazine articles along with direct observation, and combine it with interview data from participants. Interview participants are invited to tell the story of their experience, with those chosen representing different levels within the same organization or different perspectives of the same situation. The objective is to obtain multiple perspectives of a single organization, situation, event, or process at a point in time or over a period of time. The research problem is usually a “how and why” problem. Case study methodology, or the case analysis or case write-up, can be used to understand business processes. Action research is designed to address complex, practical problems about which little is known. It involves brainstorming, followed by sequential trial-and-error attempts until desired results are achieved. 中央資管：范錚強

Triangulation: Merging Qualitative and Quantitative Conduct studies simultaneously Ongoing qualitative with multiple waves of quantitative Perform series: Qualitative, Quantitative, Qualitative Quantitative precedes Qualitative Triangulation is the combining of several qualitative methods or combining qualitative with quantitative methods. This slide identifies the four strategies for combining methods. This is a good place to show the DVD case on the development of the Lexus SC 430. A significant part of the research used in the development of this sports car used qualitative research and the video describes Qual-Quant clinics. 中央資管：范錚強