once you have your statistical hypothesis then you need to think about how you will design your research in fact setting up statistical hypothesis is closely related to the research design as operationalization is all about the research design first and foremost a research project must be feasible within the available resources such as time and money so you should be able to complete and finish your project given the time and money therefore there are lots of things you need to consider for example how many subjects or participants you need to recruit for the project in general the more you have the better your outcome will be but you know you cannot have an infinite number of subjects so you need to decide how many will be enough for your project and who will be your subject humans or animals normals or patients your hypothesis will also play a role in selecting the type of subject a measurement is directly related to the data collection and for a quality data collection you need to choose carefully what you measure is what you intend to measure there can be a multiple ways to or equipment to measure the same construct for example to measure visual acuity there can be many different ways to measure it depending upon the purpose of your research if your research involves human subject then you need to go through the ethics board when there are special population or patients then you need to take extra steps to ensure their well-being finally you need to think about the actual design of your research research design can be categorized in different ways based on the different aspects of research so for example a research can be mainly qualitative or quantitative depending upon the nature of data collected so if you intend to collect the data in the form of words or media then this kind of research can be categorized as qualitative research and most of the qualitative research usually explorative and descriptive on the other hand if the data collected in the form of numbers then this kind of research is categorized as quantitative research and the data collected in this form are subject to further numerical analysis as opposed to qualitative research design on the other hand a research can be categorized into either descriptive or analytic research depending upon the goal of the study or the purpose of the study so if you are mainly interested in measuring and summarizing the pattern or frequency of a variable of interest then this kind of research is categorized as descriptive research design on the other hand if you're interested in the dynamics between the variables based on the descriptive data collected on those variables then this type of research is categorized as analytic design so in analytic design hypothesis about the variables are formally tested to find out the nature of the dynamics between those variables a research can also be categorized into observational or experimental based on the role of the investigator in an observational study the investigators act as innocent bystanders watching measuring or investigating the variables of interest without attempting to change any of those variables under investigation on the other hand the investigator in the experimental design will be actively involved in manipulating one or more exploratory variables to see their effects on the other response variables this design is almost always analytical where hypotheses about the relationship between the variables are formally tested in fact an observational study can be an analytic at the same time too however when a notable relationship is observed between the variables of both observational and experimental design it is only the experimental design to be able to say something about the causal relationship between the variables assuming that the experimental study is carefully controlled however you cannot suggest this kind of causation from the seemingly same result of the observational study because of several reasons that we will see later in this class so before we talk about those differences in design between the correlational observational and experimental design let's talk about why a simple correlation between the two variables does not necessarily mean they are causally related because we make this false inference quite frequently for now just remember that correlation between two variables does not provide a sufficient evidence for them to be causally related even though strong correlation is a necessary condition for a causal relationship so you may heard that playing violent video games or watching violent tv makes children violent let's assume that you want to test this claim on your own and found very strong correlation between the two variables where playing violent video game is the exploratory variable and the violent personality or tendency or behaviors measured by frequency of violent behaviors or languages is the response variable of the study so your study suggests that the more you play the violin video games the more violent you become assuming that the study was very well managed and designed and the evidence was legit then what do you think would people take away from this study so this is the hypothetical result on the horizontal axis we have our explanatory variable a which is the the number of hours playing violin video game and on the vertical axis we have the response variable b which is the violin behavior as the data suggests the more you play the more violent behaviors are displayed and this relationship is almost perfectly correlated in this case can we conclude that the original hypothesis is supported meaning that violent video game is a true culprit behind the violent behavior well it may be or maybe not you know when two variables change together or they are correlated with each other their relationship is not necessarily causal for example it is possible that event a is the direct cause of b if this is to be the case then we can say that the original hypothesis is supported where playing violent video game is the direct cause behind the violent behaviors however we will see exactly the same relationship when b causes a 2. if that is the case then what that means is that violent people are attracted to play more violent video games not the other way around so our assumption about the original hypothesis is completely reversed in this case so if the reverse is a possibility then these two variables can be perpetually causing each other so in other words the more you play violin video game the more violent person you become the more violent person you become then the more you will seek out to play more violent video games so i hope now you see where this is going to take a message here is that you cannot make a definite conclusion regarding the existence or the direction of a cause and effect relationship only from the fact that a and b are correlated there can be two other possibilities behind the correlation and we will take a look at different examples in the next slide okay so the data in red are based on the officer statistics from the u. s office of management and budget whereas the data in black are from the centers for disease control and prevention of the us so what's on the graph is the yearly change in the u.
s spending on science space and technology and the number of suicides eye hanging strangulation and suffocation the yearly changes in these two variables are plotted together on a double y-axis where the left-hand red represents the u. s spending and the right and black represents the number of suicides as it is obvious two lines are running almost perfectly parallel so what that means is that they increases they increase hand in hand as years go by so the more the us government spends on science space and tech the more people commit suicide oh wait maybe it's the other way around no matter how strange it may seem the correlation between the two events is so strong that it is close to almost 100 percent now do you think that this is a meaningful correlation at all you don't think that u.
