Niloy Sikder When writing up my Ph.D thesis my strategy was to always be doing something. If I got stuck on some point, I'd construct a drawing (it was all Rotring pens and pads in those days) or work on an item for the appendix. One can't afford simply to ponder something for days. So, my view is if you have a set of headings (the usual: background/intro/Literature search; Methods & materials/instrumentation; Results; Discussion; Conclusion; Appendix) or similar (let's not argue about these as it's not my main point), then always be writing these in parallel. Always try to be doing something that moves you forward. Expand your headings into sub-headings and a skeleton thesis emerges.

Dear Niloy Sikder

Please see a similar thread:

https://www.researchgate.net/post/What_is_an_ideal_approach_to_start_writing_different_sections_of_research_papers_Which_section_should_be_written_first_and_which_should_be_later

Methods and results because these can change dramatically during the research as you undertake the field work

The post that you have recommended is very good. If I am aware of something when consulting other theses, it is that each person does it in a different, perhaps personal way, and I understand that there will be structures that benefit more types of theses than others. This is also important to keep in mind, although it is clear that following the most accepted structure is a sure bet.

Hi,

The regular order of titles and subtitles of a scientific research paper, at least in the social sciences, is as follows:

A. introduction

B. Literature review

C. Research question (and possibly derived hypotheses).

D. methodology

E. Results

F. Discussion

G. Conclusions.

Niloy Sikder, I think that a couple of responses above here have gone off the track in that they deal with the structure in which to finally present research rather than the sequence in which to write it.

So, back to your question: I think that it's best to write all sections simultaneously but with some sections shaping up to their final form to a greater extent initially. I recommend those "early" sections, be the method and results sections, particularly the method because you know what you've done and getting it behind you might give you a sense of accomplishment.

Then, I would work on the introduction and discussion sections, probably largely in tandem with each other, even though they will already have some writing in them from earlier on. I recommend that the introduction is probably the last section to polish. After all, as the cat said to Alice in Alice in Wonderland, you don't know where you're going until you've been there. That might even apply most to the abstract.

Through all of that, I recommend that you work on the references. They tend to be a bit "lightweight", so when your brain isn't operating effectively with more difficult aspects, you can move ahead with the references - unless you do something else totally for a while. That's sometimes the best idea.

Happy writing!

Thanks Robert Trevethan for pointing it out. I have made the topic more clear in the description.

I did some more reading including the previous thread that Nyawira Mwangi mentioned, and decided to follow this order:

Results > Methodology > Introduction > Literature Review > Discussions > Conclusion > Abstract > Title

Please give you opinion on the pros and cons of composing a paper in this sequence.

Niloy Sikder, your choice of sequence seems fine to me, with the proviso that I think most sections require constant bits of polishing at the same time as attention is directed primarily to other sections. So, in my view, it's not a good idea to think that one section can be finalised first, then another section worked on and finalised, and so on.

In my view, a polished manuscript needs to go through a number of iterations, often with what happens in one section influencing improvements in other sections.

I'm sorry I have to disagree a little with Robert on this. How can you conduct good research and collect results, without having a clear rational for the study that is based on the existing literature and/or theory?

In my opinion, collecting resorces and studying the existing literature should be the absolute beginning and foundation of your research from which you build. If you do not have a clear picture in your head of how everything 'fits together' (the bigger picture) then how can you apply anything to your research post-partum?

At the very least you need to identify where your research sits within the existing literature AND have your hypotheses in place, which are derrived from the literature/theory BEFORE you conduct it, so that you do not end up missing something or needlessly conducting research that doesn't actually contribute anything. If you conduct the research first, then you will have a really hard time writing the literature review as you will be looking for research that supports your findings or explaining their non-significance, rather than forming hypotheses that are derrived from the research. You must have hypotheses based upon the existing literature and/or theory BEFORE conducting it! This is very important. A solid literature review will, in the majority of cases or more, result in consideration of other variables and testing parameters that you may not have even considered on your own. It will shine light on areas where your study might be improved or could even end up in a complete rebuild or even entirely different research. The literature review is there to help you explore your research question. Without it, you are stumbling in the dark.

Conducting the research first is very poor scientific proceedure and would be generally frowned upon. The most common way that research is conducted is as follows...

1. You have an idea/theory that you think would be interesting to test, and based on your own thoughts, you think you might have an idea of what the answer is, or maybe you don't but want to explore it.

2. You go and search for literature and/or theory on the subject to find out what is already known about it, which MASSIVELY impacts upon your whole study. For example, if you have a research question that has no literature AT ALL but find studies that may be related to the phenomenon, then you may decide to conduct a multiple/hierarchical regression analysis to explore how well certain variables account for the variance that explains the phenomenon (the DV) that you are intending to explore. Without studying the literature you may have conducted an entirely different experiment and one that may have been innappropriate to answer your research question.

3. After conducting the literature review you will have a clearer idea of which variables might account for the phenomenon that you wish to investigate, which might contribute the most, and how you might test with them, which allows you to then...

4. ...FORMULATE & STATE YOUR HYPOTHESES / research question - without your literature review you can not claim to have any testable hypotheses, as you have no scientific backing to do so, and, speaking from a resarch perspective, it is unwise to do so.

N.B. - it's like saying ''I believe that pollen makes people happier, because when there is high pollen, people seem to be more positive... so, i'm going to test the relationship between pollen count and people's happiness'' wellbeing/health/happiness of countries that get more sun per year and thos that get less > S.A.D. > vitamin D production and the sun > lack of sun and lower vitmain D > overcast vs sunny > hypotheses.') - Collect losts of sources and be reflective throughout the entire process. Keep yourself focused and on track, and try not to forget your research question or detract from it too much. Compare studies against one another and critique them. What did each study do well, and what did they miss? This may influence your experimental 'design' that you report in your 'methods' section.

• C. Research question (and possibly derived hypotheses) - (Split the rest of the wordcount how you see fit from here)

The research hypotheses come IMMEDIATELY after your literature review, some include a short paragraph before them with a brief summary of the identified hypotheses before stating them clearly and individually.

• D. methodology - From here on you are really writing these sections AFTER you have conducted the experiment, other than the 'analysis plan.' You should remain in past tense, where possible ie. talk about 'what you did/have done' not 'what you will do.'

Briefly describe the 'measures' (sub-heading) used and their reliability (conduct a cronbach's alpha test from your respondents, if you can, to give an idea of the appropriateness of your measures, after collecting your data). Describe the characteristics of your 'sample' (sub-heading) who were the participants? How many males vs females? How old were they? Where were they from? and identify any issues that may influence the interpretation of your results e.g. is it a representative sample? Can you generalise from the sample to the population, or are there limitations? (i.e. are they all university students from the same university? How might this impact your inturpretation?). Also include where your data was from (did you collect it yourself, or are you conducting a secondary analysis from a pre-existing data set?). How was the data collected? Opportunity sampling? Stratified sampling? etc? Finally outline your analysis plan (sub-heading). How are you going to analyse the data and what inferrential statistical tests will you use?

• E. Results

Describe the data – does it meet the assumptions of the statistical tests/analysis you propose to use? (in general, a statement of assumptions being met or not met is sufficient, but for undergraduate/post-graduate I would recommend briefly explaining how you observed each assumption being met to show understanding of the concepts e.g. by explaining how you dealt with any outliers (malholanobis distance, cook's distance, centered leverage values?) or non-normality (transformations of data, or non-parametric tests [broadly speaking transformations > non-parametric analysis]). Follow this up with the criteria you used for exclusion of certain results/participants? Justify why you excluded them? Present the results from your inferential analyses, include correlation tables, regression tables, graphs that illustrate important points, where appropriate (Don't copy straight out of SPSS, R, or any other stats programs! Make sure they are all presented in-line with your governing body/format as each has different requirements, e.g. APA vs Harvard, its not just referencing thats different tables/graphs must be presented a certain way too!)

• F. Discussion followed by a concluding statement

Discuss your findings, their wider implications and possible weaknesses of the dataset, future implications, why your results might be important, generalise to the population as a whole if appropriate, offer explanations for any non-significant results, refer to research earlier identified in the literature review etc. that supports your findings and explain how your research contributes to the field, etc. Above all REFLECT, REFLECT, REFLECT, you must always think with a critical mind. This we get you the best marks.

Sorry, it's been a long one, but I hope that helps. I wish someone had set it out for me like that when I began as an undergraduate student. It would have really helped! Best of luck with your project.

Alistair Page, with respect, I wonder whether you have missed the point - or maybe I have. At least as I understand it, Niloy was asking about the sequence he might follow when writing his research up, not when conceiving of, and conducting his research.

Your account would be very useful for many people who are embarking on a research venture. But I'd also note that your account is very much in the mould of the experimental / hypothetico-deductive frame that might not fit a lot of research. For that kind of research, however (exptal, h-d) your account would, indeed, be very useful.

I would add that, although the kind of preparation that you describe is essential, my experience is that things often alter along the way in real-life investigations and, therefore, preparing a full (and final) literature review before embarking on data collection is likely to be misguided. My experience has often been that, having had a pretty clear initial idea of the directions I wanted to pursue (as per your recommendations), I needed to go back and alter my account of those directions near the end of writing up in order to achieve a fully integrated manuscript.

That has quite a lot to do with Abraham Kaplan's distinction between logic in use and reconstructed logic.

I would add that I think most aspects of a manuscript can do with constant tweaking as the sections are being polished into a final state.

But maybe I've pretty much misinterpreted Niloy's question. If so, I hope that both you and Niloy will forgive me. :-)

Robert Trevethan - (oops, used the test account so reposting) - Hi Robert, I think that I may have misinterpreted the original question, also! After reading it again, I can entirely see where you were coming from! So, I offer my apologies. Either way, I believe your answer to be suitable for those who have already conducted the research and mine for those in preparation. My background is in the social sciences and psychology as a whole. So, May suffer a little bias toward those fields :) - best wishes.

Dear Aa Shore, your message is sweet. Thanks. I guess that, at times, we all misread questions (and, Heaven forbid, we misread people) - and it's good to be big enough to acknowledge our having done so when it happens. Good on you.

My background is psych, too, though for the final years of my university time I taught research methods. Well, we psychs are sort of forced to rub our noses in that kind of thing, particularly stats, so it wasn't too far from home for me.

So, anyway, that means I can share your bias. I reckon it's not a bad one to have.

All the best with your work.

Alistair Page - hello Alistair. Wow, I got into a tangle, or was it you? Let's say it was both of us. As you'll see from above, I responded to an Aa Shore, which seems to be an "alias" for you. I think I gave you the best I could by way of appropriate response up there. Ha ha, but I really appreciated the double apology.

Cheers, and all the best. Ah, but given that we're both in psych, I think we should acknowledge that psychology often leans toward quantitative research (and that's my personal bent), but there is a lot of excellent qualitative research - and some research of the kind characterised by "we thought we knew where we were going, but ,wow, look where it's taken us".

Again, all the best.

Everyone, apologies for the ambiguity.

I think I need to rewire the topic of this discussion, AGAIN!

Hello Niloy, maybe give it a go, then, and see whether we can help you with regard to your new topic. I think that there is some confusion about whether you are writing an article or a thesis (you refer to sections as well as chapters) - though I think it's pretty clear that the issue is writing things up, not conceiving of and conducting the study. But am I barking up the wrong tree?

Niloy Sikder When writing up my Ph.D thesis my strategy was to always be doing something. If I got stuck on some point, I'd construct a drawing (it was all Rotring pens and pads in those days) or work on an item for the appendix. One can't afford simply to ponder something for days. So, my view is if you have a set of headings (the usual: background/intro/Literature search; Methods & materials/instrumentation; Results; Discussion; Conclusion; Appendix) or similar (let's not argue about these as it's not my main point), then always be writing these in parallel. Always try to be doing something that moves you forward. Expand your headings into sub-headings and a skeleton thesis emerges.

i think first intro then literature than background . methods& materials developed automatically if u have strong literature review

Hello, Greetings all,

I read all the replies before mine. Really appreciated all the researcher's valuable opinion and thanks to Niloy Sikder for issuing the topic. By the way sorry to answer your asking in a different manner which will be very effective for you (my personal opinion).

Inasmuch research means "search again".

So you should research your own queries by yourself. Start your study. Many questions may charge your mind. Then ask specific issues like; which needs to draft first either RO or RQ? How theories will support your research? What would be the social implication of your research? Or even how to draft a good working/draft title? and many more.

And, research has no specific guidelines to maintain any heading/sub-heading order (As per my experience).

I certainly recommend you to read Alistair Page answer with great concentration.

Kindest Regards.

Rahaman, M. S.

First of all title of the article (Research Problem) should be clear, then Literature Review based on the research problem should be done and Introduction written, then Materials and Methods, Results, Discussion, Conclusions and at last "Abstract" of the article should be written.