There is nothing intrinsically deductive, or inductive, about most social research methods. They are deductive when they are used to test hypotheses derived from an existing theory and inductive when data is collected in order to develop a theory. Surveys can and are used to do both; and sometimes neither. Having fixed questions and response formats does limit the ability to use surveys inductively. Interviews and focus groups can also be used inductively or deductively , though they are more commonly used inductively. Some research approaches are inherently inductive, e.g. ethnography, grounded research; though I have seen ethnography used in an attempt to test particular, high-level theories (Structural-Functionalism, marxism, structuralism, etc.).

To my knowledge, survey research is an application of inductive reasoning. The aim of surveys is to generalize the results from a sample to a larger population (statistical generalization). It is for this reason that large sample size in surveys enable plausible generalizations. In many cases, the survey tools include few open ended questions which seek full-length responses from the participants. Depending upon the amount of data, you can use content analysis, which can result in quantitative data, or you can use thematic analysis. If the qualitative data is quite a lot, then it would be appropriate to call it a convergent mixed methods design. However, it depends on your research question. For purely quantitative survey, qualitative data can not be used.

Surveys are most commonly used to test hypotheses, which makes them deductive research (i.e., they start with theories that determine the appropriate observations).

Testing hypotheses requires the "operationalization" of concepts into measurable variables. In other words, you need to have questions in your survey that correspond to the theoretical concepts you want to capture in the form of empirical data.

Many surveys do include a view open-ended questions that produce qualitative data, but the core of survey research relies on questions that are asked in the same way to every respondent, who replies by choosing from a predetermined set of response options (i.e., yes or no, strongly agree through strongly disagree, etc.). This is very different from qualitative research, where participants are free to answer the question in whatever way best expresses their own opinions.

David L Morgan is right. I should correct myself here. Deductive reasoning is associated with quantitative research and inductive is for qualitative research.

Survey researches are purely deductive researches mostly performed from questionnaires pre-determined responses

There is nothing intrinsically deductive, or inductive, about most social research methods. They are deductive when they are used to test hypotheses derived from an existing theory and inductive when data is collected in order to develop a theory. Surveys can and are used to do both; and sometimes neither. Having fixed questions and response formats does limit the ability to use surveys inductively. Interviews and focus groups can also be used inductively or deductively , though they are more commonly used inductively. Some research approaches are inherently inductive, e.g. ethnography, grounded research; though I have seen ethnography used in an attempt to test particular, high-level theories (Structural-Functionalism, marxism, structuralism, etc.).

I agree with David Morgan's response to the second part of your question. The problem/value of open-ended questions in surveys is that you can't probe or test the responses, or ask for examples. The nature of surveys (unless they are conducted by interview of course) makes it hard to explore the meaning of answers.

Survey research method which took on a form of questionnaire (closed questions) is a deductive reasoning approach. However, in addition to a number of closed questions, you may want to design your questionnaire model flexible enough to your respondents to let them raise some issues as well for possible future line of investigation (inductive reasoning).

So in this type of flexible questionnaire model your survey research can collect some information in the form of qualitative data as well in conjunction with quantitative data.

As a result, it would be obvious that, the scores between the two sets of data might slightly be different from each other and emphasized more on the quantitative part for the convenience of respondents (deductive).

Now, it's up to your purpose of research, you may want used deductive approach to analyze the quantitative data of your survey questionnaire or inductive approach to analyze the qualitative information as the result of the flexible questionnaire model. Next, integrate them in the interpretation of results to either support or oppose your findings.

The hypothetico-deductive frame is dominant but not absolute. Complex phenomenon appeal to case study, which frame is poorly appreciated. Otherwise, think about the dynamics that these frames should represent. Don’t be troubled by the inductiv-deductive dichotomy per se. Visualize the space of your phenomenon. Ask, what is happening? Which frame can illuminate the space with rigour. Note, when you choose one frame, the other remains available as a tool...oops, you might say, too complex!

similarly, the qualitative-quantitative dichotomy is troubling.

Credit in this string: See the ‘most‘ in the reply of David L Morgan

The first part of your question has been appropriately answered by other colleagues (deductive). I would like to contribute to the second part. Having open-ended questions in a survey serves a good purpose - provide further explanation. It does not necessarily make it inductive or deductive but depends on what you choose to do with the data to generate insight. You can use a deductive approach to re-code the qualitative data in a way that allows you to analyse quantitatively or you may choose to leave them as qualitative comments and allow the ideas to develop inductively. However, you can also deductively create themes from the qualitative data based on your theoretical context in analysing the data

David L Morgan thanks for your specific response. I have an additional query regarding my own research. I am writing a protocol for a survey based quantitative study where all questions are close-ended. Is it mandatory to mention theory in the study proposal? What can be ideal explanation if I don't mention any theory in the protocol?

Thanks in advance

Abu-Hena Mostofa Kamal If you are testing hypotheses, then you should mention the sources of those hypotheses. Alternatively, if your work is purely descriptive, then you do not need any hypotheses or theories.