Loading
Notes
Study Reminders
Support
Text Version

Field Surveys and Inventories

Set your study reminders

We will email you at these times to remind you to study.
  • Monday

    -

    7am

    +

    Tuesday

    -

    7am

    +

    Wednesday

    -

    7am

    +

    Thursday

    -

    7am

    +

    Friday

    -

    7am

    +

    Saturday

    -

    7am

    +

    Sunday

    -

    7am

    +

In today's lesson, we will study about survey research methods.I have titled it as field surveys or inventories.Now surveys are one of the most important methods of inquiry as far as development research is concerned.If we look at the definitive ways in which survey research methods are studied in general in the field of research methods, it is very extensive and it covers a large number of materials.However, for the purpose of today's lesson, particularly suiting to the needs of development researches and development research methods, there are three important things that I would like to cover, which is very frequently used in development research.So, what are these three things?First, we will look at what is survey research, what is survey design, what do we mean when we are saying that a survey needs to be designed and what are the few things that needs to be kept in mind; very general points with regard to the design aspect of it.We will also cover about sampling.The most important methods of sampling that are executed as far as development research methods is concerned.Finally, I will end today's lesson with field surveys and inventories particularly looking at the characteristic features of field mapping and field inventories.Looking at some of the examples and how those examples can be best suited to the needs of the research investigation and development research methods.So, let me begin with what is survey research.Now if we look at a brief background of which are those social sciences that have extensively used survey research, we would see that the major impetus for development and refinement of survey methods came from the fields of sociology and political science.But over time, various other disciplines within the social sciences have also embraced survey research methods as their own.For example, anthropology and economics are two important social sciences which make extensive use of survey research methods.But if go to the roots of when survey research methods began, we see a lot of its use within the disciplines of sociology and political science.Actually, resistance to survey methods came from the disciplines of economics and psychology.However, with time practitioners, administrators and researchers in the field seem to have caught on with the idea of survey research.And now, there is a long history of survey research methods owing to its widespread use.Whenever we talk about development research, particularly with regard to project evaluation or evaluation studies, survey is one of the most important techniques or methods of inquiry that is carried out.However, survey methods are still not very well understood and properly executed, and that is primarily because of not understanding the technicalities with regard to survey design and sampling methods, and that is something that we will touch upon in our lesson today.Now conducting a rigorous survey research is more complicated than it appears and many researchers actually make very fundamental errors in study design and sampling procedures.And this adversely affects the instrument that is being prepared to be able to carry out the survey, particularly with regard to questionnaire design.Although questionnaire design and the important aspects to keep in mind for questionnaire design is not being taken as a part of today's lesson.However, some of the important points with regard to the designing aspect of it can goon to inform better designing with respect to survey instruments or questionnaires.Now what is survey research?Some writers define surveys as a method of data collection or tool to aid in gathering information.And in this view survey research is a broader term designating all methods of research that incorporate any type of survey and with time and with the different methods of survey being carried out online as well as offline, face to face interviews, different methods of being able to carry out the survey has also come up like we have telephone interviews,we have web surveys and so on and so forth.Therefore, the nature of survey research has become very varied.However, the very basic considerations with regard to the design and sampling procedures have not changed much.Now others bypass definitional issues completely and focused on types and methods.And there is also a view that while survey research has much in common with other research methods, particularly at the design and data analysis stages, there is a special character or logic to survey research methods that make them unique and that warrant consideration of survey methods.McGraw and Watson 1976, which is a very acceptable definition of survey research, they define it as ëa method of collecting standardized information by interviewing a sample representative of some population.í And notice the terms here ëinterviewing a sample, which is representative of some populationí.This is precisely one of the main reasons why survey research methods are carried out in development research.The idea being to support representativeness of the data that we are collecting, representativeness of the population that we are doing our research on.Now given the very basic nature of how survey research design takes place, it will not be wrong to say that survey research or survey techniques follow the overall paradigm of positive methods of experimental methods as we now call them.However, understand that the mixed methods paradigm or the pragmatic paradigm also consider survey research methods is one of the important methods of inquiry.A useful way to begin looking at survey research method would be this.First, we need to have a choice of a survey design.Next, we move on to the selection of a representative sample and choosing a sample is of utmost importance because the question here is whether we want the sample to be representative of the population or not.And often when we are carrying out surveys, representativeness is possibly the most important thing that we are trying to focus on.Then of course, interviews, questionnaire design and administration or administering of the questionnaires, how we want to administer them, whether we want to do it physically by carrying out face to face interviews, or we want to carry out telephone interviews for online surveys and so on.Now the major disciplines today that make frequent use of survey research methods are the following; sociology, political science, psychology, economics, social anthropology,education, social work, public health, and medicine.Also, in the fields of evaluation studies falling under each of these themes that Ihave just cited, survey research methods are used quite frequently.Now, time is a very important dimension with regard to quantitative research and that helps us to identify, which are the kinds of method of inquiry that should be carried out and usually survey research methods or surveys are carried out with respect to cross sectional studies or longitudinal studies.So, we can broadly divide survey research methodsí application to studies in the form of cross-sectional studies and longitudinal studies.So cross-sectional studies are basically a point of time study.You are carrying out a study at a point in time, whereas longitudinal studies take place over a period of time.So, under cross-sectional studies, what we do is observations of a sample or cross section of a population are made at one point in time and most surveys usually use cross-section data.And some of the limitations with respect to survey data making use of cross-section data also comes from the fact that the social processes social phenomenon is not best explained with the help of cross-section data.So, these limitations are pointed out.As a result of which most evaluation studies make use of cross-section data quite often.So, this leads to a common criticism of survey research that it is ahistorical and unsuited to the examination of social processes.Second is longitudinal studies and these studies permit observations of the same population or phenomena over an extended period of time and therefore, it is said that longitudinal studies enable analysis of change.That they may also facilitate more credible assertions relating to causality.What causes, which are those variables that cause a certain change in the model that we are trying to investigate.So, it is said that, the survey research methods can bring out best results based upon the longitudinal studies.Now there are three very popular longitudinal studies that make use of survey research.One is the trend studies, cohort studies, and panel studies.Trend studies basically examine change within a population over time.For example, we have the census of India surveys carried out decennially or we also have theNational Sample Survey Organization surveys which are carried out every 5 years or quinquennially.And similarly, we have various other surveys for example, agricultural census of India Or various other labor bureau studies that are carried out, education surveys that are carried out that take place over a period of time.So, in this case the sample that may be studied is not the same over a period of time.However, that gives us a sense of a pattern change that is happening over a period of time.So, there are trend studies there are also cohort studies.And these cohort studies examine over time specific subpopulations or cohorts.They may not necessarily be the same, the individuals may be different.For example, we may study or interview people aged 30 years in 1970 and then 40 years in1980, 50 years in 1990, 60 in 2000 and so on and so forth.So, we are studying cohorts at different points in time.And that also gives us some kind of longitudinal data.However, the technique or the method of inquiry that is being carried out is survey research.A third kind of study with respect to longitudinal research studies are panel studies.These basically examine the same set of people each time.Example, when we are interviewing the same sample of potential voters every month during an election campaign.So, the cohort studies and panel studies differ with respect to the population that we are interviewing.The cohort studies may interview similar groups of population, but the individuals may not be the same.Whereas in panel study, we are investigating the same set of population over a period of time.So, and the method of inquiry is again the survey research methods.So, these are the few aspects to keep in mind with respect to the time dimension of sampling consequent upon the survey design that we want to carry out.Based upon the discussion that I have had so far; we can have three types of surveys.One is self-administered questionnaires, self-administered surveys wherein the interviewer approaches the interviewee personally.There is a physical interview.It can be either physical interview or we are self-administering the questionnaire through a certain medium.It may be email surveys, web-based surveys or group surveys in a classroom.Now there is a slight bit of a difference made between questionnaires and schedules.Usually schedules are questionnaires made where the interviewer himself or herself physically approaches the interviewees for writing down the details in the schedule.Whereas questionnaires are usually close ended and does not leave much room for interpretation by the interviewees, and it is self-administered, but it can be sent across with instructions to the interviewees to be filled up by the interviewees themselves.And then we can also have telephone service.So, let us look at some of the advantages and disadvantages with regard to these different kinds of surveys.Self-completion, telephone survey, and face-to-face interview.Now the advantages of self-administered surveys are one is that it is cheap.The interviewer does not need to be present.We can cover wide areas.The anonymity of the interviewee is protected.And there are no interviewer bias and people can take their own time for completing the survey questionnaires.The disadvantages are that because the interviewer is not present for a face-to-face interview,and the questionnaires are being sent across either through emails or through group-based surveys, the response rate may be very low.We will come to how response rate is calculated in some time.And possibly there may also be a bias with respect to the responses that we are receiving from the interviewees.The disadvantages, questions need to be very simple because there is no scope for the interviewer to simplify the questions for the interviewees when they are filling up the questionnaires.There is no control over interpretation and there is no control over who fills it.And of course, with respect to self-administered questionnaires or the right term to use is actually self-completion of the questionnaires.Because questionnaires are usually administered by the investigator or the interviewer, and it is completed by the interviewees.Usually, self-completed questionnaires need a lot of reminders from the investigators to the interviewees to be sent back to them.So, some of the tips that one should remember when one is focusing on self-completed questionnaires is that the questions must be very simple or it must be simplified and not leave much room for interpretation.There must be a cover letter explaining the intent and motivation of carrying out the survey and possibly also giving instructions with regard to the surveys that we are conducting.Include stamped addressed response envelope and also keep sending reminders.Telephone surveyís advantages are that we can do it all from one place.We can clarify answers because we are speaking to the interviewer ourselves.People may be relatively happy to talk on the phone because it also enables some kind of anonymity on the part of the interviewee.It is also relatively cheap and quick.And the disadvantages are that people may not have phones.You may get a wrong person or call at a wrong time.There may be a bias from whose name is listed or who is at home, and it is easy for people to break off and there is no context to interview.So, what are the tips to remember with regard to telephone surveys?You rely totally on verbal communication.Therefore, questions must be short and words easy to pronounce, and to minimize number of response category- so people can remember them because you are asking them.Usually telephone surveys options could be either in the form of a yes or a no.But when we have more options or more categories to choose from, it becomes difficult for people to remember.Face-to-face interviews are the methods of inquiry that are most often used in development research methods, because when we are moving out to the field to carry out the kind of research questions that I have discussed over the period of last few weeks, these are the questions that are best investigated when the investigator is present on the field and therefore face-to-face interviews is something which is very extensively used in development research methods.And there are some very robust advantages with regard to these interviews.First is that you have a very high response rate and you are in control of the interview situation and you are also able to clarify responses.So, the respondent bias is reduced as far as face-to-face interviews are concerned.Disadvantage is it may be very expensive.It may be very time taking.It is a resource intensive exercise.There is a lot of time and money spent on being able to carry out face-to-face interviews.The interviewerís presence may also influence a lot of questions that are to be answered,which is also called the investigatorsí bias.So, both the respondent bias and investigator bias are frequent.Or the incidence of this is very high when we have face-to-face interviews.And if there is more than one interviewer, they may have different effects.Suppose usually when we carry out surveys in rural areas or urban areas, we carry out household surveys, door to door surveys and we have different teams of investigator approaching these households with the same survey instrument, the same questionnaire.The responses from the interviewers may be varied based upon the kind of questions that are being put forth by the investigator.So that also affects the quality of the responses that is been received.Some of the tips to remember with regard to face to face interviews are that the interviewer should be non-threatening and therefore, usually when we go for development surveys in different locations, we always give an introduction of ourselves pointing out to the intent of the research, where it is being used, whether we are trying to publish it, who are we collaborated with, with respect to the dissemination results and so on.So, all of this information is made very clear to the interviewee to begin with.Therefore, the interviewer should be non-threatening.The interviewer can clarify questions, but should be wary of elaborations that affect the content.Usually when we explain more than required to the interviewees the kind of answers responses that we are getting also gets affected.So, we have to apply some amount of judgment with regard to how much should be discussed with the interviewees.We need to ask questions in a very clear and standardized way and if the list of possible responses is long, we need to show them to the respondent for them to read while the question is being read out.So, these are some 10 steps to a good survey design.First, we need to identify what is it that we want to cover in a survey.Suppose we are carrying out a study on complementary feeding practices.This is the example which I have been taking in the last few weeks.And we want to investigate the particular questions with regard to food diversity, food composition, the frequency in which food is being fed, and the various socioeconomic correlates that go on to affect these outcomes, complementary feeding outcomes.However, along with asking questions and complementary feeding, if we go on to ask remotely related questions, let us say on different other kinds of nutrition outcomes with regard to the households,then it may go on to jeopardize the research investigation that we are carrying out in the first place.So therefore, we need to make it very clear with regard to what is it that we want to cover in our survey that we have chosen to do.The next step is to put questions that are necessary.The first is this identify.Second is we need to put questions that are necessary.Unnecessary questions should be removed because that affects the length of the questionnaire.And when the length of the questionnaire gets affected, it is very difficult to make the respondent sit down for a very long time and answer the question.So, the respondents may clearly be very bored with the questionnaire the length of the questionnaire that we have and we may not get the suitable answers.So therefore, there is a need to keep it short and simple.We ask one question at a time, avoid using jargons.Often when there are technical terms that we put in the questionnaires it also needs to be seen how best to deconstruct those technical jargons and ask the respondents about them.And these issues can be dealt with only when we have face-to-face interviews.And this is also one of the advantages of face-to-face interviews.Suppose we are carrying out a study on nutrition, a nutrition survey is being carried out andwe do know that there are certain technical terms that are used with regard to anthropometric measurements let us say for example.We are taking estimates on mid upper arm circumference or we are taking measurements on the body mass index, taking measurements that can inform us about stunting and wasting and underweight.Now if we start using these terminologies on the field, it is but obvious that we may not get the required answers because the respondents may not be aware of these technical terms and therefore, it is important for the investigators to simplify these terms or not use technical jargons at all.Whether to use open ended questions or closed ended question is something which needs to be deliberated upon by the investigators.Because face-to-face interviewing gives us a lot of scope to carry out investigations with the help of both open ended and close ended questions, and well-designed surveys usually contain both of them, close ended questions as well as few open-ended questions.We need to spend time to design our survey, analyze the responses after collecting, puta summary report together and conclude our plan of action.These are the 10 steps to a good survey design that needs to be kept in mind.Those who are interested in looking at more aspects of a good survey design, I would urge you to look up this website, questionpro.com from where we have retrieved this figure on10 steps to a good survey design.Now having discussed some of the very basic issues with regard to survey design or what is survey research and different kinds of surveys that are carried out.This is a very elementary level of looking at the sampling procedures.However, this has been kept at an elementary level so that development researchers who want to make use of quantitative data can begin to think about why and where which procedure of sampling is to be used.So, before we introduce some of the very common forms of sampling that are carried out in development research, let us also get clarified certain terminologies with regard to sampling procedures.First is the difference between a sample and a population.So usually when we say that we are choosing a sample that is representative of a population,what is it that we mean?Because the concept sample and the population are used in a certain manner, when we are looking at sampling techniques.So, before that what is sampling theory?Sampling theory is basically a study of relationships between a population and samples drawn from a population.So, for example, you have a class of hundred students, and you want to look at the weight and height of children in a class.And if you are looking at hundred students in a class and the total number of students in the class is hundred, then the 100 N is equal to 100 is your population.Whereas, if you want to draw sample from the class, let us say, a 30% sample from a class of hundred students to look at the various other socioeconomic characteristics of those children then what you are essentially doing is you are drawing a sample from the population.So and why do we draw a sample from this population, because if we do not have the time and resources to study all the students in a population, and we want to carry out some procedure of sampling to be able to draw a sample so that, that sample can tell us a lot about the larger population from which the sample is being drawn.So, in that sense a sampling theory is a study of relationships existing between population and samples drawn from the population.Now it is useful in estimating unknown population quantities such as population mean and variance often called population parameters.When we are trying to calculate the average weight and height of a class of hundred children,let us say, based upon the population and we are carrying out some averages there, what we are essentially doing is we are estimating population parameters and these are referred to as parameters.So, when we are calculating the average age of children in a class of hundred, so hereage becomes a parameter.If we are calculating height, height becomes a parameter and so on from a knowledge ofcorresponding sample quantity such as sample mean and variance.And this is also referred to as sample statistics.When we are carrying out these estimations based upon the sample and not the entire population then essentially what we are expressing is sample statistics.So, population parameters, we want to estimate a population parameter let us say age or height based upon the sample statistics from the sample that we have drawn from the population of hundred children.Imagine increasing this number of 100 to 1000 or 10,000 and then drawing a sample from a population of 1000 would then start making sense.For example, based upon the census data, when we calculate the average literacy rate of the country or we calculate the average life expectancy of adults in a country, so essentially what we are doing is we are giving you population parameters here.In the case of National Samples Survey Organization when we are carrying out employment statistics,we are saying what is the employment rate in the country, what is the usual status unemployment rate, what is the daily status unemployment rate.What we are essentially doing is we are carrying out samplesí statistics.We have drawn a sample based upon a sampling procedure and then we are carrying out, weare estimating sample statistics based upon the sample that we have drawn.So that is about population parameters and sample statistics.Now a range of statistical analysis of a sample can be carried out including descriptive statistics and that comes under the larger domain of development statistics.For those who are interested can look up various books related to development statistics.One of them being Chandan Mukherjeeís Econometrics in Underdeveloped Countries.This is a highly recommended book for those interested in studying development statistics that are applicable to development research methods.Now the thing to keep in mind is the topic of interest or research question is typically focused on some feature of the population.Example whether on average women in India earn more or less than men as opposed to whether the women in the sample in question earn more on average than the men.So, this is a point that needs to be remembered, what is the motivation for carrying out a survey?What is the motivation for drawing a sample from the population that we want to study?So, when we are posing a question with regard to let us say complementary feeding and we are saying that, what are the complementary feeding practices that are followed in a certain community, in a certain location, in a certain state in India let us say.So, if we want to study the different kinds of feeding practices in a certain community and the community has large numbers of people, then obviously we will be drawing the sample from them, from that population to check if we want to adopt the method of survey to do an inquiry with regard to their complementary feeding practices.So usually when we are posing the research question for which survey inquiry is to be carried out, research questions are never posed for the sample that we are studying,the research questions are posed for the population that we are studying and to be able to answer those questions we draw a sample from the population and then say that okay, since we have followed the rules of sampling for drawing a sample, so the estimates that are coming from that sample are representative of the entire population that we are studying.So just to summarize, this discussion that we have had, I will just read out the whole slide for you.Sampling theory is a study of relationships existing between a population and samples drawn from the population, number one.Number two, what are population parameters?Population parameters, it is useful in estimating unknown population quantity such as population mean and variances.These are often called population parameters or briefly the parameters from a knowledge of corresponding sample quantities.And when we are estimating sample mean and sample variance it is referred to a sample statistics or briefly statistics.And there is a range of statistical analysis of a sample that can be carried out including descriptive statistics.Descriptive statistics is usually the very basic estimations that we carry out with regard to the measures of central tendency let us say carrying out the arithmetic average- the arithmetic mean, the media, and the mode and the skewness in the data, the variance in the data and so on and so forth, the standard deviation in the distribution that we are studying and so on.Lastly the topic of interest or research question is typically focused on some feature of the population example we can ask the question whether women in India on an average earning more or less than the men and we hardly ever ask the question whether women in the sample in question are earning more or less on an average than men.Now in general, study of the inference is made concerning a population by using samples drawn from it and together with indication of accuracy of such inferences we are using probability theory and this is called statistical inference.