Typically you are conducting research for a reason - you want to know something. In this case you are testing a theory-based supposition that the intervention will or will not impact the test group. Thus you essentially have a hypothesis whether you want it or not :) 

Typically you are conducting research for a reason - you want to know something. In this case you are testing a theory-based supposition that the intervention will or will not impact the test group. Thus you essentially have a hypothesis whether you want it or not :) 

I feel to support David. As it is a theory testing case hypothesis is essential.

I teach my grad students that they MUST HAVE a 'burning question' or they will find researching any topic difficult.  In addition, to stay motivated during the difficult process of completing as research study, one must also pick a focus (a hypothesis or set of related hypotheses) that will sustain one's academic career in profitable and productive ways.  A smart approach in the research is to collect more data than you want to report in the first set of studies so you can build on initial studies and keep the field and yourself in a positive position.

Sounds like your hypothesis is either  THERE IS a difference in learning literature between the program group and the control group or THERE IS NOT a difference. So you are testing the hypothesis of either your intervention is effective or it is not effective. 

Do meu ponto de vista é possível que não haja hipótese em pesquisa quase experimental quando se trabalha com manifestações diversas [como se fossem sintomas] e que não se tem um problemas previamente estabelecido em forma da pergunta. Então, nesses casos o pesquisador, mesmo tendo um grupo de controle, sai à busca do problema [do tipo causa, não necessariamente única] das ditas manifestações. Em resumo: pesquisa que não tem um problema previamente determinado com clareza pode sair buscando tal problema e logo não tem hipótese. 

At the end of your work, even if you don't include it in your research paper, you will end up with a hypothesis. But a truly believe that, just like in the case of a action research, you can use a Investigation Question to start working upon a quasi-experimental research.

In the case of my University: Univerisdad Católica Boliviana, when you are doing a action research in the educational field, you are asked to have a Investigation Question or a investigation problem instead of a hypothesis.

A hypothesis is entirely possible and appropriate, however a hypothesis is only relevant if the conclusion explicitly and exclusively addresses the hypothesis (the conclusion must answer the hypothesis as supported or not supported by the discussion and/or evidence.

Even more important is that you start with an examination of your own beliefs and assumptions about the phenomena you want to study.  You have to consider your own biases and observations and be as transparent as possible as you explore the state of research in your area of interest, consider your philosophic framework, the theory that is in keeping with that philosophy, and how you can advance the field by posing new hypotheses to study.

If you have an experimental design with a control group and you want to test the possible differences between the two groups, you definitely have an underlying hypothesis. It seems that the aim of your research is to test the efficacy of distillation on teaching in a specific context. Hence, you are actually testing the null hypothesis : H0= there are no differences between the program and the control group vs. H1= Differences will be observed (i.e. the program group will have a better outcome).

Good luck with your research!

I support Valerie's point: when you compare two groups you will always have a hypothesis, no matter what. Why would you compare two groups if you weren't to assume that there exist differences between those groups? ;)

Of course this is a very open hypothesis; the next step would certainly be to specify on which dimensions you think the two groups will differ and whether there is previous empirical evidence that supports the assumption that one group will be better/stronger/... than the other - then you would have a directed hypothesis instead of an undirected one.

I think so.  For example, you might have a hunch that culture makes a different in learning but unsure as to how to measure that.  In that case, the hunch could posed as a question and not a formal hypothesis such as "Does culture make a difference in learning?", organize a focus group to ask that question, and then from the focus group answers, formulate a more formal hypothesis that can be measured.

 Mike

Also, having a hypothesis, or a question that you have that serves as the foundation for the research, enables you to better explain the purposes of the research to others...funders, colleagues, etc. Makes it easier for non-researchers to understand the rationale behind the research.

Yes, it is crucial to have a hypothesis in quasi experiment whether directional, non directional or null hypothesis.

Given that you wish to compare two groups you have an inherent hypothesis, or you would have no interest in the two groups. However, because this is quasi-experimental you will likely have bias, which must be controlled a-posteriori for your results to be interpret-able. To accomplish this you may wish to look at non-parametric analytical techniques because they are more forgiving.

I think it's possible, because you explain the objectives 

What makes the program group being the program group? And what makes the control group being the controll group?

If you have an answer for that (e.g. program group gets more reading lessons or reads articles instead of book or...) then you already have a hypothesis. You just haven't specified it for you yet.

I agree with most answers, and I would say just one more remark: When you experiment, it is much easier to falsify an hypothesis, than to validate any idea. Then, best experimental works serve to falsify given ideas and force you to forge some "theoretical" (i.e.: stronger and consistent) ideas.

When conducting a research, a hypothesis will help ground your research questions. More so in your study as  you are testing a theory-based supposition as to whether the intervention will or will not impact the test group. Thus, a hypothesis will surely do justice.

I think that as my colleagues above wrote, you actually have a hypothesis (that the intervention will make a positive difference). However, as a fellow literature scholar, I know the field of literature despises hypotheses and prefers to merely examine the material rather than make assumptions regarding the findings - as this is considered "tainting" the findings with your own beliefs or suppositions. But I would say that this study falls under the category of education research, where such hypotheses are common and welcomed, as long as you demonstrate that the study structure and analysis were not bias towards a certain outcome. You can also say: we tested to see whether this intervention would make a difference, and perhaps thus not phrase it as a hypothesis but rather as a research question. You can also look at what KIND of differences there were, and this would again be a research question rather than a hypothesis, with more qualitative answers (this is mostly relevant if the study has open questions).

Te recomiendo leer el libro Diseños experimentales y cuasi experimentales en la investigación social de Campbell y Stanley. 

The most important thing is to have a problem definition. The hypothesis can emerge once you've begun to  collect data. In all grounded theory research and qualitative research, you won't start with the construction of a hypthesis, and only progressive try to find out what makes "sens".

recuerde que la formulación o no de hipótesis depende del alcance de la investigación y del paradigma investigativo que pretenda acordar

We all approach his subjective research with our own views about how to deal with a study for which one has no initial hypothesis(es) as it is exploratory research in search of a direction  Of course, one must have a section that define the problem.  For me, this is problematic because the problem IS the lack of a clear sense of what is going to be found.  The argument becomes circular -- ergo, one MUST start with their own hunches and their own burning question.  Otherwise the research contributes little to the field and to the sense of accomplishment the researcher wants to get from doing the study in the first place.

There is a fundamental difference to make between burning question and hypothesis, to my opinion. Hypotheses are le centre of "classical" and experimental research designed, whereas grounded theory seeks to put in place tools which allow to construct sens during the research process. Look at the writings of Glaser, or Miles & Huberman for this. 

I agree with Oscar Fonseca: all depends on the research paradigma. As long as this is not determinated all questioning about the sense or non sense of hypothesis is useless, in my opinion.

Thanks, Monica, for your further clarification of differences between burning questions and hypotheses.  Of course, they are fundamentally different as is research vs. evaluation and the methodologies employed by each.  I teach research design to grad students and they often struggle with developing hypotheses from any kind of grounded theory.  Glaser was a colleague and mentor of mine and I use his book as one of the key contributors (along with the Miles & Huberman text)  to our current understand of the research process.  That understanding has evolved over recent years and what I teach is a way for beginning scholars to see where they fit and what paradigms appeal to them in either basic or applied research areas.  The newer paradigms for 21st century research are more collaborative, transparent, and oriented to breaking some of the classical research rules  (see work by Andy Hargreaves, Michael Fullan, Peter Senge,and Otto Scharmer).

In Cook and Campbell 1979 book "Quasi Experimentation: Design and Analysis Issues for Field Settings" discuss the "Hypothesis" in this classic book on the topic.  Their discussion implies that the hypothesis does exist in quasi experimental designs.  However, they do discuss the hypothesis in the context of threats to internal validity and that quasi experimental designs may lead to "hypothesis guessing" by the subjects because of this design. 

It is always a good idea to have a shared pre-test and post-test. At least you want to have some index that tells you the groups are comparable to start. It is also important to know whether the students in both groups interact, whether the same teacher teaches the different class and with comparable enthusiasm. Finally, you really need outcome measures.

As a journal editor I can say that not having hypotheses is going to impact you negatively during the peer review process. The result will be that many journals will reject your paper while as others that are more open minded will likely ask you to add a hypothesis that is theoretically grounded. Good luck.

I also do guest editing and sit on several editorial boards for AERA and APA journals.  Our Educational Psychology members have started innovative online journals via Frontiers in Psychology online and welcome the types of exploratory studies that have null hypotheses and are looking for new ways to examine certain phenomena and advance the field.  The turn around time to get published is much shorter and us editors have strict guidelines for evaluating research from worldwide scholars who are actually reading our US research!

Toda investigación cuasi-experimental requiere de hipótesis que al final puedan probar las preguntas propuestas en el problema

A research is always conducted to investigate an issue or a question in hand having an underlying theoretical construct. When the research design consists of a Control and an Experimental group and the investigation is to find the impact of research intervention, a hypothesis is definitely required. The interpretation of results will reveal whether the hypothesis is proved positive, negative or is it null hypothesis. 

At the very least you have as a hypotheses- The use of distillation has a measurable impact on learning. Students who are exposed to a literature lesson using a distillation methodology perform better on a post test than students receiving a traditional instructional methodology. Of course you would also examine the null hypotheses. Additionally, you might examine the subcomponents of the distillation methodology.

Nancy,

This may be of help: From UNICEF

"Main points Quasi-experimental research designs, like experimental designs, test causal hypotheses. A quasi-experimental design by definition lacks random assignment. Quasi-experimental designs identify a comparison group that is as similar as possible to the treatment group in terms of baseline (pre-intervention) characteristics. There are different techniques for creating a valid comparison group such as regression discontinuity design (RDD) and propensity score matching (PSM)."

Quasi-experimental design is a special case of experimental design. The difference is the absence of randomness. Basically, hypothesis is

Indague Libro Metodología de la Investigación. 5t a edición. Le dará esa respuesta.

A quasi-experimental design is considered an experimental design but it lacks either a control group (as in a time-series design) or random selection (as in a non-equivalent control group design). The quasi-experimental design was developed by Campbell and Stanley to provide an alternative to the constraints educators have when trying to do experimental research in classroom settings. In such a design you are always testing a hypothesis with the intent to draw causal conclusions.

It may not be possible to draw conclusions in a casual effect comparison without testing the hypothesis

Nancy,

No, quasi-experimental designs are used to explicitly test hypotheses. They are called "quasi" because true randomization is not possible and/or there is no comparison group.

It depends on what is meant by "quasi".

A single Subject Psychophysical study might generate data to compare with a math or statistical model. Indeed, all model testing studies may, or may no, have more than one condition -- where values or n of parameters are varied.

There is a different kind of research that might be so labelled: Assessment of a school treatment variable. It is usually the case that the assumed independent groups or schools change during the study. Principals leave, students are absent, the 'control' school gets wind of what is happening in the 'experimental school'; pupils get reassigned. All normal happenings, i.e. one cannot count on a fixed variable value assignment. What to do?

- Watch very carefully and treat each group as different learning conditions, being prepared to track students' learning in a matter of fact way, e.g., reviews, homework, etc. Somehow find a way to monitor big changes, e.g., what is on the walls, notebooks of both teachers and pupils. Rochel Gelman