greetings everyone last week Sweden introduced the notion of appalled and a sewer wall I noticed in your comments to the forum that many of you were frustrated with a number of pseudopod so-called faith walls and many of you appear to have dismissed the polling often the other basically were not able to articulate the case for legitimate same difficulty valid Paul and it's utility in a modern society and this is the kind of topic that I would like to discuss this week to clarify some of these concepts and basically possible to clarify exactly what a good scientific Paul is as you correctly pointed out there are meaning to the Paul's fundraising under the guise of Pauling selling under the guise of Pauline the so-called pushpots that we'll discuss later in the semester also there are lots of different threats to validity of scientific pawn and we'll discuss some of these threats in the following weeks for example many of the Paul's have inappropriate objective basically we call these series cereal balls for instance some organizations as matter of fact a lack interest in measuring people's preferences instead they're trying to push a particular agenda also oftentimes the interest in surveying or gauging the public's opinion may be there but the methodology is incorrect for instance the sampling procedures may be flawed or the questioners themselves may be ineffective as some of you pointed out it's not uncommon to see the so called bloated language in the questions and also some Paul's some surveys amid key questions for example they fail to collect information about people's democratic with their characteristics as a result ESO is a pretty useless for scientific research or for in the context of a more thorough investigation so what exactly are the attributes of a good scientific Paul it is very complex issue and that's the kind of issue that we'll discuss in the following couple of months and it also will cover different strategies for amplifying for enhancing the validity of surveys my gf diffuse to make sure that you can recognize problems with studies and also that if every you need to put together your own survey you know exactly what to do how to come up how to end up with an excellent valid survey the kind of survey from which you can eventually draw good inferences so let's define what a survey is definitely there are three different definitions some of which you can find and think one of the book that we're using this semester not definition guys you can find in Asha well this is my definition which summarizes what this college did much of the scholarship in the field of research methodology understands to be an essential property of a good soviet it is a systematic method for gathering information from individuals in order to identify their preferences and other attributes and also typically the purpose of generalizing to the broader population as you can see I have underscored some of the important characteristics so a good survey once discussed first of all the good Sylvie is characterized by a so-called systematic approach it's not some haphazard methodology and that's exactly how many people view the Paul's and it's just a one time questionnaire not really guided by any specific criteria well there are different guidelines and you have to follow these guidelines in order to be in a position to draw valid inference will discuss these guidelines in the form of weeks so systematic approach guys remember it's a very important characteristic you cannot just go with it for me can't just go outside and stop asking people questions even if you have an attempt of engaging people's interests and preferences it will not really end up with a good survey as a result another important characteristic of a survey methodology is that we target individuals in the social sciences they're different what we call units of analysis you may be interested in an organization potentially you might be interested in comparing countries and we call this field comparative politics there's a field that's the field of international relations serving methodology is unique in that it's individuals individual human beings respondents that's your into analysis of course in the process we need to identify what we call a sample and a property sample in a couple of weeks we'll discuss different challenges of sample once again you need to be quite methodical and systematic when you consider which individuals to select for your survey in terms of attributes what exactly we're trying to measure here so these are pretty unique as a method object because they allow us to measure people's values attitudes preferences they also collect information about people's demographic characteristics there are other ways to measure preferences and values and I'll discuss one for you some of these alternatives a little bit later but basically pâtisserie methodology is one with the best and probably most effective ways to measure attitudes any challenge you guys try to come out right to think of an alternative way and ultimately this is the main utility of Pawlik several of you pointed that out there's no there's super no other way for us to understand human behavior to measure both behavior and the preferences that drives a variety of behavioral responses another very important part of a good survey it's different from pseudo Paul's a good survey is a kind of service from which we can extrapolate to a broader population let me define what a population is now it's in inductive statistics the essence of induction as a mode of reasoning it's to look at a small sample and from that sample we try to make inferences that are applicable in a much broader context and we call that population there's no way you can survey all 300 million Americans well I guess theoretically is possible but this would be very costly process our understanding of statistics allows us to select more sample of Americans and by looking at the characteristics of these people then we'll be able to generalize to the entire country of course the same logic works in other settings as well that's what we call the generalization extra plating sample results to a broader population from which a particular sample has been drawn and there's a property of both surveys all the survey's differ because of their scope and right now let's discuss briefly guys different types of surveys so that you can recognize that there's a variety of different survey the doll G by the way once again in your comments on the forum most of you pointed out or most of you ended up focusing on political so they post the wax or pre-election type of surveys but the servant methodology is very diverse there are different types of surveys and many of these service will qualify as valid types of Paula so let's discuss some of some of these different types of surveys in terms of instrumentation some surveys are self administered think about self-administered questionnaires something that you can receive online or by regular mail and then respondents expected the feeling these response forms whatever xx feedback forms a good example would be class evaluation for all no one really ministers that the students expected to log on to particular University website and then to feel our evaluation for an alternative approach it is interviewing there's actually someone a professional who administers with a survey but the cost and benefits of each approach I discussed and think and actually refers to that as well so I would prefer not to dwell this particular issue right now but you need to recognize once again that in terms of instrumentations rather two fundamentally different approaches although in some cases a hybrid approach is used as well in which these fermenters are the people who use service combined both self administered question other question is and in the element of interview in terms of the scope we need to recognize that sometimes what we call a sample at service sample basically encompasses pretty much the entire population think of a census of course there's a trade-off the broader your sample is the logic it is in terms of the number of people involved what were the last questions you will be able to ask so when we think while the census once again we use that as an illustration typical people are asked only few questions that's it otherwise the costs will be established they're quite significant anyways at times the entire group the population is really small what they used the same example of class evaluation pose it is quite possible and feasible to ask every student in class what he or she thinks about the course about the instruct about it materials about the assignments so this was an example of a survey that encompasses that involves incorrect in scientific surveys it is more common to deal with you to attempt to describe the properties of a very large population in these instances in terms of the costs typically the the researchers collect a very small sample a random sample of a large population and that's the kind of surveys that probably we'll focus on in the rest of the semester scientific balls think about Gallup polls 1,000 individuals can be randomly selected from 300 million Americans and then there's survey the key or good scientific poll is that you can generalize from this sample to a broader population in terms of purpose who if you go back to think she discusses different purposes of surveys sometimes certain surveys are very broad in terms of the types of information that the researchers are trying to collect for instance national election studies that are administered on a regular basis every couple years typically in the Indy surveys and poor well I'll introduce you to some of such surveys you can see that they include hundreds of questions they deal with different aspects of people's preferences and attitudes in contrast there are some stories that are very specific in terms of their objectives once again US Census only asks a few questions few basic demographic facts about the population that's all that we're interested in in an author that that's all that we can afford to collect also another example would be survey of students preferences in class when would you prefer to present your research results now this would also be a survey very now in terms of its scope also quite narrow in terms of its purpose in the form by the way you suggested some excellent examples after the survey is quite narrow in terms of their purpose the purpose will still be quite legitimate for instance the so-called straw polls in Iowa and which the Republicans were selecting or expressing their preferences for presidential candidates technically not all ballots are also types of a survey with a very specific mission to allow people to select representative these Paul's these types of surveys definitely do not collect demographic information about the people and Hillary rarely do they ask other questions about people's social economic or political preferences only in this mess will also discuss different types of design for scientific surveys and you guys won't need to understand the differences and also the structure the costs and benefits of different designs right now please allow me to use just a couple of terms cross-sectional surveys or cross-sectional design and longitudinal surveys the cross-sectional designs are very common and this is surveys that collect information at a single point in time national election science 2004 2010 or General Social Survey as a 2006 would be examples of cross-section designs in contrast many surveys and many research projects are longitudinal in nature these properties tend to be much more expensive that's downside of the surveys but they allow you to collect information over a period of time so we call them longitudinal and this is the kind of methodology that also allows us to track the evolution of changes in the population over time so definitely it's it's a very important type of a design so let's go back to my original question are there any alternatives to surveys I definitely want you guys to think about this question this was my question as a matter of fact that I posed to you in a slightly different form that we could go I serve these essential can they be replaced as a methodology as a method for collecting information about people's preferences well I came up with my own ways of alternatives okay yes spend a couple of hours I can send it all different options and well he's not list performance test absolutely you can use something like that to measure someone's attitude or someone's skills abilities but of course we need to understand that performance test think about different standardized tests that you were exposed to they quite limited in applicability they're very costly they're time consuming and ultimately I would question their utility because we measure people's political social or even economic preferences by offering different tests to individual but that's a quite effective if if you're interested in measuring summers academic scholarly or professional potential well sounds physical capacity capabilities definitely the performance test may be quite effective but if you're interested in understanding someone's political views it turns out that the performance tests have very limited utility okay what else do we have experimentation in many Sciences for instance in biology physics chemistry in psychology to some extent the researchers can experiment the essence of experimentation is that you have two different groups of individuals or whatever else is your unit of analysis you apply a particular treatment to one group of course you assign the subjects randomly to these different groups and you withhold that treatment from the second group possibly you expose it or you give it to what we call placebo a fake treatment and then you compare the results that's what we call pure scientific experiment unfortunately experimentation is impossible virtually impossible in the context of the social sciences particularly if you were interested with in measuring the properties of aggregates of human beings for instance organizations or societies or countries but even with human beings guys there are simply too many ethical concerns ethical questions experimentation is in quite impressed as himself we don't use experimentation a lot well what is left is what I would call the observational approach now let me tell you it is quite common in the fields of anthropology in political science as well well in sociology possibly to less extent in psychology and the economics definitely observations are less intrusive by the way that was one of the objections of many of you in your comments on the form that the polling itself can be very intrusive and it's time consuming and that's what people don't like well an alternative would be simply to observe people and attempt to collect information about their preferences they bought yourself going through a village in Bangladesh or some way in the Amazon well but you could go to Paris France and you simply observe the locals and you take very thorough notes and you try to determine their political preferences or their social values or their economic habits it is possible guys it's definitely an alternatives through surveys but think about the downside oh and it's quite tremendous first of all this particular approach can be less intrusive but it is very costly also it is very time-consuming its costly in terms of the resources that you spend and time is one of these resources well this knowledge itself can be quite unreliable it is not clear to what extent we can generalize from such observational studies and it is also quite difficult sometimes we decipher the records when you observe people's behavior does this behavior actually reflect people's underlying attitudes and preferences this room means to be a big issue in terms of validity of option statics guys there's no way we can measure individual human preferences directly as a result we're stuck with a survey methodology so oh and I almost forgot that there's another approach of course you could also review reading records and documents you can conduct a thorough text analysis it's also possibility it's also a possibility but there's a significant time delay for instance if you would like to know people's preferences what Americans think about the wars in Iraq and Afghanistan or people's attitudes towards for instance national health care plan where will you find such information there are no documents out there that will summarize people's values ultimately these documents emerge they will be based on some type of survey or observational methodology so text analysis sometimes it works particularly if you're interested in some type of historical investigation for a for instance political values in France in 1789 well you cannot interview annual anyone at this point what you can do is go to archives and analyze the texts available to us but when we deal with the current environment with the concentration serving methodology definitely offers probably the best way of looking into people's preferences so let's summarize what is the utility of quality after all and can we completely dispense with the power well I have to agree with some of your comments that if we remove Poli if we rely less on the modern society will unravel okay that's that's definitely it's it's Isabella point it's about a point at the same time serving methodology is really useful to us as well first of all there's one of the few ways that we have right now to measure to gauge public opinion and public preferences will discuss this issue for them in the following weeks whether the politicians should rely on such Paul's necessarily but at the same time in a modern democracy in the modern society that is important to know what the population actually wants and what the preference is in the population are it is quite important both for the rules and as several of you pointed out in your comments and is also important for the people that's one of the ways for the people to express their opinions and their preferences well think about working about for an organization surveys are quite essential when you work for a larger organization when are you planning a particular policy so these typically come quite commonly of an essential element of the so called pilot studies when you planning for a particularly event you might be interested in people's response this allows you to set up the priorities in a more effective in a more efficient way so and this is definitely one but feels value of silly methodology soubise also allow us to adjust the existing policy class evaluation form they used for you to express your assessment about a particular University course but also it allows the instructors to change things to tinker with the material with the schedule and with the course of syllabus surveys are very essential they provide the necessary feedback for the decision makers finally as I have already discussed surveys are absolutely crucial in the context of the social science this is simply one of the few methods that we have to collect information about human preferences about individual values and about individual attitudes we don't really have any other viable and cost-effective alternatives as a result surveys constitute a major methodology in statistics in statistical analysis in which the researchers are trying to uncovered different relationships between the variables