Chapter 2 Quantitative, Qualitative, and Mixed Research This chapter is our introduction to the three research methodology paradigms.

A paradigm is a perspective based on a set of assumptions, concepts, and values that are held and practiced by a community of researchers. For the most of the 20th century the quantitative paradigm was dominant. During the !"0s, the qualitative paradigm came of age as an alternative to the quantitative paradigm, and it was often conceptuali#ed as the polar opposite of quantitative research. Finally, although the modern roots of mi$ed research go bac% to the late !&0s 'and its historical roots go much further bac% in time(, ) thin% that mi$ed research truly became the legitimate third paradigm with the publication of the Handbook of Mixed Methods in Social and Behavioral Research '200*, by Tasha%%ori and Teddlie(. At the same time, mi$ed research has been conducted by practicing researchers throughout the history of research. Characteristics of the Three Research Paradigms There are currently three ma+or research paradigms in education 'and in the social and behavioral sciences(. They are quantitative research, qualitative research, and mi$ed research. ,ere are the definitions of each Quantitative research . research that relies primarily on the collection of quantitative data. '/ote that pure quantitative research will follow all of the paradigm characteristics of quantitative research shown in the left column of Table 2. . 0ee Table 2. in your te$tboo%.( Qualitative research . research that relies on the collection of qualitative data. '/ote that pure qualitative research will follow all of the paradigm characteristics of qualitative research shown in the right column of Table 2. .( Mixed research . research that involves the mi$ing of quantitative and qualitative methods or paradigm characteristics. The mi$ing of quantitative and qualitative research can ta%e many forms. )n fact, the possibilities for mi$ing are almost infinite. Quantitative Research Methods: Experimental and Nonexperimental Research The basic building bloc%s of quantitative research are variables. Variables 'something that ta%es on different values or categories( are the opposite of constants 'something that cannot vary, such as a single value or category of a variable(. 1any of the important types of variables used in quantitative research are shown, with e$amples, in Table 2.2 'see your te$tboo%(.

)n loo%ing at the table note that when we spea% of measurement, the most simple classification is between categorical and quantitative variables. As you can see, quantitative variables vary in degree or amount 'e.g., annual income( and categorical variables vary in type or %ind 'e.g., gender(. The other set of variables in the table 'under the heading role ta%en by the variable( are the %inds of variables we tal% about when e$plaining how the world operates and when we design a quantitative research study. As you can see, independent variables 'symboli#ed by 2)32( are the presumed cause of another variable. Dependent variables 'symboli#ed by 2D32( are the presumed effect or outcome. Dependent variables are influenced by one or more independent variables. 4hat is the )3 and D3 in the relationship between smo%ing and lung cancer5 '0mo%ing is the )3 and lung cancer is the D3.( 4henever you want to ma%e a claim about cause and effect 'i.e., that changes in one )3 cause changes in another )3( you have to be very careful about what are called extraneous variables 'i.e., variables that compete with the independent variable in e$plaining the outcome(. 6erhaps the D3 did not change because of the )3, but it changed because of an e$traneous variable7 8ou will learn how to 9control for: these %inds of variables in several places in your boo% 'including below when we briefly discuss e$perimental research(. 0ometimes we want to understand the process or variables through which one variable affects another variable. This brings us to the idea of intervening variables 'also called mediator or mediating variables(. )ntervening variables are variables that occur between two other variables. For e$ample, tissue damage is an intervening variable in the smo%ing and lung cancer relationship. 4e can use arrows 'which mean causes or affects( and draw the relationship that includes an intervening variable li%e this0mo%ingTissue Damage;ung <ancer. 0ometimes a relationship does not generali#e to everyone= therefore, researchers often use moderator variables to show how the relationship changes across the levels of an additional variable. For e$ample, perhaps behavioral therapy wor%s better for males and cognitive therapy wor%s better for females. )n this case, gender is the moderator variable. The relationship be type of therapy 'behavioral versus cognitive( and psychological relief is moderated by gender. /ow, ) will tal% about the ma+or types of quantitative research- e$perimental and none$perimental research. Experimental Research The purpose of e$perimental research is to study cause and effect relationships. )ts defining characteristic is active manipulation of an independent variable 'i.e., it is only in e$perimental research that 9manipulation: is present(. Also, random assignment 'which creates 2equivalent2 groups( is used in the strongest e$perimental research designs.

,ere is an e$ample of an e$periment. 6retest > > Treatment ?@ ?< 6osttest >2 >2

4here @ stands for the e$perimental group 'e.g., new teaching approach( < stands for the control or comparison group 'e.g., the old or standard teaching approach( and 2 subscripts stand for time- Atime one= 2Atime two. Because the best way to ma%e the two groups similar in the above research design is to randoml assign the participants to the e$perimental and control groups, letCs assume that we have a convenience sample of &0 people and that we randomly assign them to the two groups in our e$periment. ,ere is the logic of this e$periment. First, we made our groups appro$imately the same at the start of the study by using random assignment 'i.e., the groups are 9equated:(. 8ou pretest the participants to see how much they %now. /e$t, you manipulate the independent variable by using the new teaching approach with the e$perimental group and using the old teaching approach for the control group. /ow 'after the manipulation( you measure the participantsC %nowledge to see how much they %now after having participated in our e$periment. ;etCs say that the people in the e$perimental group show more %nowledge improvement than those in the control group. 4hat would you conclude5 )n this case, we can conclude that there is a causal relationship between the )3, teaching method, and the D3, %nowledge, and specifically we can conclude that the new teaching approach is better than the old teaching approach. 1a%e sense5 /ow, letCs say that in the above e$periment we could not use random assignment to equate our groups. ;etCs say that, instead, we had our best teacher '1rs. 0mith( use the new teaching approach with her students in her &th period class and we had a newer and less e$perienced teacher '1r. Turner( use the old teaching approach with his &th period class. ;etCs again say that the e$perimental group did better than the control group. Do you see any problems with claiming that the reason for the difference between the two groups is because of the teaching method5 The problem is that there are alternative e$planations. First, perhaps the difference is because 1rs. 0mith is the better teacher. 0econd, perhaps 1rs. 0mith had the smarter students 'remember the students were not randomly assignment to the two groups= instead, we used two intact classrooms(. 4e have a name for the problems +ust mentioned. )t is the problem of alternative e$planations. )n particular, it is very possible that the difference we saw between the two groups was due to variables other than the )3. )n particular, the difference might have been due to the teacher '1rs. 0mith vs 1r. Turner( or to the )D levels of the groups 'perhaps 1rs. 0mithCs students had higher )Ds than 1r. 0mithCs students( 4e have a special name for these %inds of variables. They are called e$traneous variable.

)t is important to remember the definition of an extraneous variable because e$traneous variables can destroy the integrity of a research study that claims to show a cause and effect relationship. An extraneous variable is a variable that may compete with the independent variable in e$plaining the outcome. Eemember this, if you are ever interested in identifying cause and effect relationships you must always determine whether there are any e$traneous variables you need to worry about. )f an e$traneous variable really is the reason for an outcome 'rather than the )3( then we sometimes li%e to call it a confounding variable because it has confused or confounded the relationship we are interested in. Nonexperimental Research Eemember that the defining characteristic of e$perimental research was manipulation of the )3. 4ell, in none$perimental research there is no manipulation of the independent variable. There also is no random assignment of participants to groups. 4hat this means is that if you ever see a relationship between two variables in none$perimental research you cannot +ump to a conclusion of cause and effect because there will be too many other alternative e$planations for the relationship. )n the chapter, we ma%e a distinction between two e$amples of none$perimental research. )n the 2basic case2 of causal!comparative research, there is one categorical )3 and one quantitative D3. @$ample- Fender ')3( and class performance 'D3(. 8ou would loo% for the relationship by comparing the male and female average performance levels. )n the simple case of correlational research, there is one quantitative )3 and one quantitative D3. @$ample- 0elfGesteem ')3( and class performance 'D3(. 8ou would loo% for the relationship by calculating the correlation coefficient. The correlation coefficient is a number that varies between . and H , and 0 stands for no relationship. The farther the number is from 0, the stronger the relationship. )f the sign of the correlation coefficient is positive 'e.g., H.I&( then you have a positive correlation, which means the two variables move in the same direction 'as one variable increases, so does the other variable(. @ducation level and annual income are positively correlated 'i.e., the higher the education, the higher the annual income(. )f the sign of the correlation coefficient is negative 'e.g., G.J ( then you have a negative correlation, which means the two variables move in opposite directions 'as one variable increases, the other decreases(. 0mo%ing and life e$pectancy are negatively correlated 'i.e., the higher the smo%ing, the lower the life e$pectancy(. 4e will show you how to improve on the two basic none$perimental designs in later chapters, but for now, please remember these important points-

( 8ou can obtain much stronger evidence for causality from e$perimental research than from none$perimental research 'e.g., a strong e$periment is better than causalG comparative and correlation research(. 2( 8ou cannot conclude that a relationship is causal when you only have one )3 and one D3 in none$perimental research 'without controls(. Therefore, the basic cases of both causalGcomparative and correlation research are severely flawed7 *( )n later chapters we e$plain three necessary conditions for causality 'relationship, temporal order, and lac% of alternative e$planations( or a previe! of these three necessary conditions re"uired to ma#e a firm statement of cause and effect, read this next section$ %t is provided as supplemental or previe! material for this topic !hich occurs in man& chapters of the 'oo#$ %f &ou have had enough for no!, (ust s#ip to the next section of this lecture entitled Qualitative Research$ There are three necessary conditions that you must establish whenever you want to conclude that a relationship is causal. They are shown in Table *. from a later chapter in your te$tboo%. >ur e$periment met these criteria quite nicely. That is, we had a relationship between teaching method and %nowledge= the manipulation occurred before the posttest= and because we randomly assigned the people to the two groups, there should be no other variables that can e$plain away the relationship. >n the other hand, in the basic cases of causalGcomparative and correlational research, where we only observed a relationship between two variables 'we had no manipulation or random assignment(, we have only established condition . 4e can only conclude that the two variables are related. )n chapter we will show you how to design none$perimental research that performs better than the basic cases on the three above conditions. 0till, remember, even when these basic cases are improved, e$perimental research with random assignment is better for studying cause and effect than none$perimental research. Another way of saying this is, if you want to show that one thing causes another thing, then, if it is feasible, you will want to <>/DK<T A/ @?6@E)1@/T. Qualitative Research Methods 4e described the ma+or characteristics of qualitative research earlier, in Table 2. . There are five ma+or types of qualitative research- phenomenology, ethnography, case study research, grounded theory, and historical research. All of the approaches are similar in that they are qualitative approaches. @ach approach, however, has some distinct characteristics and tends to have its own roots and following.

,ere are the definitions and an e$ample of the different types of qualitative research "henomenolog . a form of qualitative research in which the researcher attempts to understand how one or more individuals e$perience a phenomenon. For e$ample, you might interview 20 widows and as% them to describe their e$periences of the deaths of their husbands. #thnograph . is the form of qualitative research that focuses on describing the culture of a group of people. /ote that a culture is the shared attitudes, values, norms, practices, language, and material things of a group of people. For an e$ample of an ethnography, you might decide to go and live in a 1ohaw% communities and study the culture and their educational practices. $ase stud research . is a form of qualitative research that is focused on providing a detailed account of one or more cases. For an e$ample, you might study a classroom that was given a new curriculum for technology use. %rounded theor . is a qualitative approach to generating and developing a theory from data that the researcher collects. For an e$ample, you might collect data from parents who have pulled their children out of public schools and develop a theory to e$plain how and why this phenomenon occurs, ultimately developing a theory of school pullGout. Historical research . research about events that occurred in the past. An e$ample, you might study the use of corporeal punishment in schools in the !th century. Mixed Research Mixed research is a general type of research 'itCs one of the three paradigms( in which quantitative and qualitative methods, techniques, or other paradigm characteristics are mi$ed in one overall study. @arlier we showed it ma+or characteristics of mi$ed research in Table 2. . The )dvantages of Mixed Research First of all, we advocate the use of mi$ed research when it is feasible. 4e are e$cited about this movement in educational research and believe it will help qualitative and quantitative researchers to get along better and, more importantly, it will promote the conduct of e$cellent educational research. 6erhaps the ma+or goal for researcher who design and conduct mi$ed research is to follow the fundamental principle of mixed research. According to this principle, the researcher should mi$ quantitative and qualitative research methods, procedures, and paradigm characteristics in a way that the resulting mi$ture or combination has complementary strengths and nonoverlapping wea%nesses. The e$amples +ust listed for mi$ed method and mi$ed model research can be viewed as following this principle. <an you see how5 ,ere is a metaphor for thin%ing about mi$ed research- <onstruct one fish net out of several fish nets that have holes in them by laying them on top of one another. The 2new2 net will not have any holes in it. The use of multiple methods or approaches to research wor%s the same way.

4hen different approaches are used to focus on the same phenomenon and they provide the same result, you have 2corroboration2 which means you have superior evidence for the result. >ther important reasons for doing mi$ed research are to complement one set of results with another, to e$pand a set of results, or to discover something that would have been missed if only a quantitative or a qualitative approach had been used. 0ome researchers li%e to conduct mi$ed research in a single study, and this is what is truly called mixed research. ,owever, it is interesting to note that virtually all research literatures would be mi$ed at the aggregate level, even if no single researcher uses mi$ed research. ThatLs because there will usually be some quantitative and some qualitative research studies in a research literature. *ur Research T&polog&

4e have now covered the essentials of the three research paradigms and their subtypes. To put it all together, view the picture of our research typology in Figure 2.M.