Study unit 7:Data collection

Question 1

how is data collected in research of social sciences

Answer 1

Measurement is a central part of analysing survey questionnaires, testing hypotheses and theoretical assumptions, and summarising data obtained in field research. Measurement is when we record (or register) the value of a particular variable. This enables us to describe the variable.

Nominal level of measurement
The categories used in the nominal levels of measurement are (1) discrete,
A category is discrete when it is distinct and separate from other categories
(2) mutually exclusive
Categories are mutually exclusive (with the emphasis on exclusive) when they measure something which is not measured by other categories.
(3) exhaustive
Categories are exhaustive if they make provision for all possible responses, variables or attributes.

Question 2

what is Ordinal level of measurement

Answer 2

Ordinal levels of measurement are used when a researcher wants to put his or her observations or data in a certain order of importance

the categories used in ordinal scales must allow data to be ranked. that the distance (i.e. interval) between categories ranked next to each other are not equal.

Question 3

what are Interval level of measurement

Answer 3

The characteristics of categories used in an interval level of measurement also correspond with those used in nominal and ordinal levels of measurement. However, unlike the nominal or ordinal scales, the interval tscale (as the name indicates) can measure the interval or distance between wo points on a quantitative instrument (eg a thermometer)

Question 4

what is Ratio level of measurement

Answer 4

The ratio level is regarded as the highest level of measurement, because the characteristics of its categories include all those we have discussed so far for the nominal, ordinal and interval levels of measurement. In addition, a ratio level of measurement has an absolute zero point. The ratio scale is normally used when variables such as weight, length and time are measured (quantified); the ratio scale is not normally used to measure people’s attitudes or perceptions.

Question 5

discuss measurement errors

Answer 5

measurement error: an inaccuracy that influences the measurement of an attribute

systematic error: bias that occurs constantly, and which distorts the measurement in a particular direction

random error: haphazard errors due to different factors, which may cancel each other out after a number of cases/measurements

1. some respondents are tired or ill
2. some respondents find the presence of the researcher threatening
3. the researcher has a momentary lapse in concentration
4. the research equipment (eg tape recorder) malfunctions

Question 6

what are The three advantages to collecting and using existing data:

Answer 6

ƒ ƒThe method is unobtrusive. (Also, because respondents are not directly involved or required to respond, any measurement the researcher is using is not influenced in any way.)
ƒ ƒThe method is practical. (Takes less time and costs less.)
ƒ ƒThe accuracy of measurements (reliability and validity) can be confirmed (because the data to be analysed concerns past events).

Question 7

what are some sources of data

Answer 7

primary data: information collected at the original source

secondary data: information taken from a source that was not involved in the original investigation

tertiary data: interpretations based on secondary sources

Question 8

what is survey research

Answer 8

Survey research can be defined as a systematic collection of specific information, obtained through asking a sample of respondents the same questions at a particular point in time. Survey research involves investigating a large number of people who are geographically spread out over a wide geographical area

Question 9

what is field research

Answer 9

Field research involves observing events and people (individuals, groups, organisations, families) in their natural settings

Question 10

explain interviews as data collection method

Answer 10

An interview, as a data-collection method, uses personal contact and interaction between an interviewer and an interviewee (respondent). Such personal contact takes place either in a face-to-face situation or via a telephone.

Structured and unstructured interviews depends on:
1. the researcher’s knowledge and familiarity of the topic
2. the purpose of the interview (eg whether to confirm assumptions, or collect new information)
3. the nature and sensitivity of the topic (eg Aids)
4. the actual respondent(s) (eg children as rape victims)
5. the setting (eg a busy shopping centre)
6. the relation between the interviewer and interviewee (eg a psychologist and a patient)

structured interview, specific lists of questions are asked
An unstructured interview allows respondents to give their reactions to general issues, in the absence of specific questions.

Question 11

what are in-depth interviews and focus group interviews

Answer 11

In-depth interviews have also been called “intensive interviews”, “unstructured, conversational interviews”, “ethnographic interviews” and “focused interviews” In-depth interviews differ from face-to-face interviews conducted in a survey, of a large population, because the main aim of in-depth interviews is to obtain detailed information

Focus group interviews (also called group interviews) consist of 6 to 12 people who are interviewed together at the same time. The interview is normally semi-structured or totally unstructured and guided by a discussion leader or moderator.
a respondent’s verbal responses may be less honest, and therefore less reliable, during a group interview. Respondents may be influenced by others and may feel the need to conform. Secondly, it may not always be possible to protect a respondent’s confidentiality or prevent the adverse effects group participation may have on certain individuals

Question 12

what are the types of questions used in interviews

Answer 12

A closed-ended question consists of a question with a fixed number of answers.

The purpose of open questions (in interviews and in questionnaires) is to encourage respondents to express their attitudes, emotions, ideas, sentiments, suggestions or opinions in their own words

Free-narration questions are also open questions, but here the respondent is invited to tell his or her “own story”.

A role-playing question is another variation of an open question. Here the respondent is invited to play a particular role and respond from the perspective of that role.

Follow-up questions are unique to interviews (and cannot be asked in postal questionnaires), because they follow on from the respondent’s answers to a closed-ended or open question

Question 13

list the advantages of interviews

Answer 13

Advantages of interviews
ƒ ƒBecause interviews are flexible (especially semi-structured and unstructured interviews) they can provide a researcher with detailed and fresh information the researcher may not have predicted or anticipated.
ƒ ƒBecause an interview involves dialogue between interviewer and respondent, unclear questions can be clarified.
ƒ ƒAdditional information can be obtained by asking follow-up questions, especially where responses are ambiguous or unclear.
ƒ ƒAn interview allows the researcher to establish rapport, or even a personal relationship, with respondents. This, in turn, enables the researcher to contact respondents at a later stage if further questions arise during the interpretations and analysis of the data.
ƒ ƒRespondents do not have to be literate.
ƒ ƒNonverbal communication (eg facial expressions) can be observed and noted.
ƒ ƒRandom samples can be drawn.
ƒ ƒThe way in which questions are asked can be standardised.

Question 14

what are some limitations involving interviews

Answer 14

Bias is a central problem. An interviewer may hold a biased opinion towards the interviewee because of the respondent’s demographics (eg race, gender, age). This can then bias the results

face-to-face interviews are costly, because interviewers’ salaries, as well as their traveling and accommodation expenses, have to be paid. Costs are also higher than using a telephone interview, because of the time it takes to prepare for the interview, to travel, to actually conduct face-to-face interviews and to transcribe recordings or preparing systematic field notes after the interview.

If several interviewers are used, this is not only expensive, but also requires training in skills — in other words, the interviewer needs to know how to ask questions, how to listen, write notes and whether he or she should anticipate the need to ask probing questions

breadth is often sacrificed for depth. Although obtaining in-depth information is an advantage of face-to-face interviews, researchers — owing to the time and expenses involved — cannot interview a large sample.

Question 15

what is a questionnaire

Answer 15

A questionnaire is a printed document that contains instructions, questions and statements that are compiled to obtain answers from respondents. The questionnaire differs from interview schedules or interview guides, because respondents fill in the questionnaire without the researcher’s assistance. This is known as a self-administered questionnaire

Question 16

what should be avoided in a questionnaire

Answer 16

ƒAvoid using the words “or”, “and”, “therefore”, “either” or “both”, because all of these words lead to statements or questions that are double-barreled. For example: “Do you agree with the increase in the medical scheme payments and hospital tariffs?” Even if a respondent answers “yes”, we are unable to determine whether the response refers to the “increase in the medical scheme payments”, or to the “increase in hospital tariffs”, or both. In effect, this is two questions joined together in one question.
ƒ ƒAvoid using a question or statement which conveys a position for or against the topic being investigated (called loaded language). For example: “Do you also prefer working flexi-hours because you can plan your workday according to your personal needs?” The use of “also” and reference to an advantage, loads a “yes” justification onto this question.
ƒ ƒLoaded language, in which phrases such as “Do you also prefer …”; “Do you agree with …”; or “Do you support the view that …” are used also result in leading questions or statements. This type of question or statement must be avoided, because a respondent may feel pressurised to agree, and this will give the researcher a biased response.
ƒ ƒIn addition to avoiding loaded language and leading questions, words such as “can” or “might” should also be avoided, because they create questions to which only one logical answer (“yes”) is possible. For example: “Can educational television programmes increase your children’s learning skills?”
ƒ ƒNegative items should be avoided, because respondents can easily “misread” and therefore misinterpret the item. For example: “Should police officers in the charge office not fulfill a public relations function for the South African Police Service?” A respondent can easily miss the “not” in this sort of question and answer “yes” (thinking that she supports the suggestion that police officers should fulfill a public relations function). A negative can in fact be interpreted positively — therefore giving us responses that contain error.
ƒ ƒ“Should we spend more on social welfare?” is an example of an incomplete question. This type of question must be avoided, because a respondent is forced to give an opinion about something that has been inadequately defined. In this case, the respondent does not know which aspects of social welfare, or which social welfare facilities are being referred to. Nor does the respondent know what is meant by “more”. More than what? More than on other services (eg education, housing)?
ƒ ƒReferences to vague agents of action must also be avoided, because they result in vague responses. In the last example above, it is not clear who or which institution is represented by “we”. Vague and imprecise responses also occur if the questions or statements are incomplete or ambiguous. For example, “What do you think about the day care facilities?” This question does not specify “what” aspect of the day care facilities we are being asked about or whether we are being asked about day care facilities on a national, provincial or local level. The question does not even tell us what type of day care facilities it means — for babies and children? Street children? People who are mentally retarded? People who are physically disabled?
ƒ ƒLengthy questions and statements must always be avoided, because they tend to be complex and obscure what is being asked or stated. For example: “Should the state raise funds to counteract poverty and also introduce a national development agency?”
ƒ ƒWe should always avoid making unwarranted assumptions in the questions or statements which we formulate, since they can annoy respondents who may not hold these assumptions. This problem can be avoided by using filter and contingency questions — discussed below in section 7.6.2.
ƒ ƒAbbreviations (or acronyms), jargon or slang must also be avoided, because some respondents will not know what they mean. For example, “AA” is an abbreviation for “Automobile Association” and “Alcoholics Anonymous”. In Britain it is an abbreviation used to classify films that are suitable for persons older than 14 years.

Question 17

what to include in questionnaires

Answer 17

ƒUse everyday language, not research or other technical jargon.
ƒ ƒUse unambiguous questions or statements.
ƒ ƒAsk only one question at a time.

Question 18

what are some items found in questionnaires

Answer 18

items in a questionnaire can consist of statements and questions which can be open or closed-ended.Closed-ended questions, which consist of two or more responses, have also been called standardised questions, fixed-alternative questions and highly structured questions

paired-comparison question. In this type of question, the respondent also only selects one option, but has to actually think about the options and compare their meanings

Contingency questions are questions that only apply to some respondents and must therefore include clear instructions

Ranking questions are also closed-ended questions, but instead of selecting one option, respondents are required to rank or order the options from the least to the most (or from the most to the least), according to their preferences.

Inventory questions are also closed-ended questions, but again the respondent is not limited to selecting only one option. The purpose of inventory questions is to obtain a comprehensive overview of all possible options that could apply to each respondent

Multiple-choice questions can be worded in different ways, depending on what is being investigated. They do share one characteristic, namely, that the respondent must select one of the options from those given. Multiple-choice questions are also often used to collect demographic data

Question 19

what are likert scales

Answer 19

Likert scale by formulating a number of statements related to the attitude or behaviour that is to be measured. In some cases, we can expect positive or favourable responses, while in others we can expect negative or unfavourable responses. Respondents are asked to read the statements and to select one of the categories (eg Disagree) which best represents their attitudes or opinions

Question 20

what is a semantic differential scale

Answer 20

Semantic differential scales consist of a series of seven-point spaces which have opposite adjectives or adverbs (referred to as bipolar ratings) on either end of each scale

Semantic
differential scales are flexible and versatile because they can be constructed to include a variety of bipolar options, thus enabling respondents to give their rating of one or several concepts

response bias (i.e. to prevent a respondent from marking all the same spaces below one another). This type of response bias is also known as a response set, whereby respondents fall into the pattern of answering questions in a particular way

Question 21

how can questionnaire items be ordered

Answer 21

ƒBegin with general or broader questions and proceed to more precise or specific items (called the funnel technique or funnel pattern).
ƒ ƒThe grouping of items should follow a logical pattern. The above guideline represents such a pattern, in which the general questions also function as warm-up questions. If we are interested in a topic that involves the time when something happened, our questions should first deal with the past, then the present and then the future.
ƒ ƒTo avoid making the respondent feel hostile, begin with neutral questions and place sensitive questions in the middle

Question 22

explain the reliability property of scales

Answer 22

reliability is the degree to which a scale yields consistent results or scores. Consistency in results does not imply that the results remain the same. If circumstances have changed, the measuring instrument should reflect this. Three empirical methods can be applied to a scale to estimate its reliability.
the test-retest method
The test-retest method involves correlating one scale to the same group of respondents at different times, with intervals in between that vary from a week to a year. The reliability is estimated by correlating the two sets of scores

the split-half method
When applying the split-half method, the scale is split into two halves, with each half scored separately. The two halves can be constructed by selecting items randomly, or by combining all the even-numbered items in one half and all the odd-numbered items in the other. The reliability is estimated according to the correlation between the scores of the two halves
The equivalent-forms method (also called alternate-forms method) involves compiling two forms (or measures) that contain items which measure the same phenomenon. These two separate forms are administered to the same group and the two sets of scores are correlated to estimate the level of reliability

equivalent-forms method.
the equivalent-forms method, you use two parallel versions of the same measuring instrument and thus estimate the equivalency component of the instrument’s reliability

Question 23

what conditions are necessary for scientific observation

Answer 23

1. First, the observation must be guided by a specific purpose, which should be specified by means of hypotheses and/or research questions
2. Secondly, observations must be related to, and therefore guided by, scientific theories. (For example, observations of a family watching television soaps, can be related to the uses and gratifications theories of mass communication, which claim that television viewers make a conscious and motivated choice among television stations and different programmes.)
3. Thirdly, the procedures to be followed during the observation(s) must be systematically planned.
4. Fourthly, the actual procedures followed and any deviations from the original plan must be systematically recorded.
5. Fifthly, where possible, the observations must be checked

Question 24

what three factors influence researchers’ observations

Answer 24

These three factors deal with selective attention, bias and the researcher’s intentions

Question 25

what is participant observation

Answer 25

the researcher (as observer) joins and becomes involved with the group being observed.

Question 26

what is nonparticipant observation

Answer 26

In nonparticipant observation, the researcher fulfils the role as an observer and is entirely removed from the social interaction he or she is observing

Question 27

what is ethnographic research

Answer 27

Ethnographic research is a specific type of participant-observer research, because the researcher’s aim here is to describe a particular group’s way of life, from the group’s point of view in its own cultural setting

Question 28

explain audience ethnography

Answer 28

The aim of audience ethnography is to describe how audiences (eg radio listeners or television viewers) experience radio or television programmes in the social context of everyday life

Question 29

explain street ethnography

Answer 29

Street ethnography takes place in certain settings, such as a street, shelter, hostel or boarding house and is aimed at describing the way of life of certain groups of people who spend their lives in these settings, such as sex workers, drug dealers or street children

Question 30

mention four categories of observation in field research

Answer 30

1. Observing external or physical signs. For example, types of clothing, colour combination of clothing and jewellery can function as signs of people’s occupations, lifestyles and income levels. Beards, and even hair styles, may signify people’s emotions or state of mind, whereas tattoos tend to indicate that a person belongs to a subculture. By observing external or physical signs, researchers observe current behaviour, and can also infer past behaviour. For example, certain tattoos may indicate that the person was once pressurised into conforming with group norms.
2. Observing kinesics. Kinesics refers to the use of facial expressions, gestures and bodily positions (postures) as nonverbal communication. Facial expressions may communicate one of many emotions, such as happiness, surprise, fear, anger, sadness, disgust, contempt, interest, bewilderment or determination. Gestures can be a substitute for words (eg nonverbal signs for “OK”, “come here”, “go away”, “be quiet”, or “I am warning you”) and can regulate a conversation between two people (eg nodding the head or folding arms whilst listening). Bodily positions (postures) can signify people’s expectations, involvement with others, and emotions, such as a child assuming a fetal position when experiencing stress. However, the meanings conveyed by kinesics, especially those that regulate conversations, are often culture-bound. Also, the meaning of kinesics may differ, depending on the particular setting in which observations are made (DeVito 1992:199–217).
3. Observing proxemics. Proxemics signify status, control or ownership in social interactions. Proxemics deals with the study of spatial communication and specifically the distances between two or more people. Hall (1959, 1963, 1966), who pioneered research into proxemics, distinguishes between four distances. Each of these four distances corresponds with four types of social relationships, namely: intimate distance, personal distance, social distance and public distance. In terms of these distances, we claim ownership of certain territories. For example, intimate distance applies when two people are touching whilst lovemaking, or comforting or protecting one another. When ownership of such a distance (or territory) is challenged, we experience this as a violation (when entering someone’s office without permission), invasion (eg parents joining teenagers in their hangout) or even contamination (eg smoking in a colleague’s office). Such encroachments prompt people to behave in different ways. For example, we may rearrange the furniture in our offices and homes to act as a barrier and to increase the distance between ourselves and others. We may use linguistic insulation by using professional jargon or a language which is not understood by others. Or we may respond by simply withdrawing from the situation. The relevance of the degree of space may differ, depending on variables such as age, gender, physical and emotional states. We can expect that proxemic rules will be violated in densely populated conditions, such as in prisons and hospitals, and this may result in the depersonalisation of individuals. (This may also explain why strangers who are forced to share a crowded lift avoid eye contact.)
4. Observing language behaviour. Apart from analysing what people say, language behaviour includes stuttering, slips of the tongue, silences, who speaks to whom, the duration of speaking, and the paralanguage dimensions of speech. Paralanguage refers to the vocal (not verbal) characteristics, such as the rate of speaking, the volume and rhythm, and vocalisations, such as “uh-uh”, “mmmmm” or “shhhh”.

Question 31

how are observations recorded In classification,

Answer 31

we assign persons and behaviours to particular categories, or we apply a scale as a measuring instrument. Such categories and scales can be devised prior to, or after, observations. Classification may in itself be the aim of our observations, or it may be used to guide and structure observationsThe labels given to categories must be operationally defined to ensure that they clearly reflect (1) what is to be observed, (2) how the variables to be observed will be measured, and (3) how the measures will be recorded.

once categories have been operationally defined, the next step is to select or devise instruments to observe, measure and analyse the data.