Write every research objective with one action verb and a clear, measurable target, and write a hypothesis only when the objective compares, relates or predicts.
Many scholars are mixing up these two. An objective tells what the study is going to do. A hypothesis is a prediction that a statistical test can check, so you write one only for an objective which a test will answer. Suppose your objective is only describing the level of something, or giving a profile of the sample. For that kind of objective no hypothesis is needed. Qualitative studies mostly work with research questions in place of hypotheses. We will use one sample study all through, first correcting its objectives and then writing its hypotheses.
Start with a weak set of objectives
Here is a sample study we made up, on social media use and sleep among college students in Trivandrum. A first draft of objectives mostly looks somewhat like this.
- To study the social media use of college students.
- To understand the effect of social media on students.
- To know about the sleep problems of students.
- To give suggestions.
Committees see lists like this many times, and their comments are mostly the same. Words like “study”, “understand” and “know about” cannot be measured, and nobody can tell when that work is finished. Objective 2 uses the word “effect”. For effect you need an experiment or data over time, and here the plan is only one survey done once. In objective 3 sleep problems are there, but where is social media? It is not linked at all. Objective 4 is not an objective at all. Suggestions come out of the findings, whatever the objectives are.
Repair them, one problem at a time
Change the verbs. In place of “study”, “understand” and “know”, use verbs which tell what you are going to do with the data, such as assess, compare, examine the relationship, identify, estimate, develop or evaluate. The verb itself gives a hint about the analysis. Our objective 2 has the word “compare”. Seeing that word, an examiner will turn pages to find a test of differences. With “examine the relationship” in the objective, they turn to the results chapter looking for correlation or regression.
Name the variables. “Social media use” is too vague as it stands. Do you mean hours per day, use after 11 pm, or the score on a published scale? “Sleep problems” can be measured as sleep quality on a standard index, for example the Pittsburgh Sleep Quality Index.
Name the people and the place. Write undergraduate students in arts and science colleges in Trivandrum, not just “students”.
Remove the objective that is not an objective. Take out “to give suggestions”. Your suggestions will still go in the last chapter.
After these changes the set becomes like this.
- To assess the level of social media use and sleep quality among undergraduate students in arts and science colleges in Trivandrum.
- To compare sleep quality between hostel and day students.
- To examine the relationship between night-time social media use and sleep quality of these students.
- To identify which factors predict poor sleep quality among these students.
Now each objective has one verb, a target you can measure and a clear analysis behind it, and you can say when it is finished.
Do one small check on your own list. Read every objective aloud and ask yourself what you will have in hand once it is done. If you will have a table, a test result or a model, that objective is fine. If you will only have “some idea” or “a better feel”, it is still loose. Correct it at this stage itself, because once the data is collected, changing an objective is lot of trouble.
How many objectives, and how they fit together
Most PhD theses have three to six objectives. If you have less than three, many times the work looks more like a paper than a thesis. If you have more than seven, usually some of them are only sub-steps of another objective, or the study is trying to cover two theses in one. When two of your objectives use the same data and the same test, ask whether they are really one objective.
The objectives should also follow one after the other in a sensible order. In our corrected set, objective 1 describes, objective 2 compares, objective 3 looks at a relationship and objective 4 predicts. This order is common and the reader can follow it easily. Describing comes first. After that you go to group differences, and only later to links and prediction of the outcome.
Also read your problem statement and objectives together. If the gap is about night-time use and sleep, at least one objective has to deal with night-time use and sleep directly. Our guide on the research gap and problem statement discuss this same link from the problem statement side.
Writing the hypotheses
A hypothesis is a prediction about the result, and a statistical test can check it. To be strict, the test only rejects the null hypothesis or fails to reject it. It never proves that the null is true. So go through your objectives one by one. Where a test will answer the objective, write a hypothesis for it, and where no test is involved, leave it without one.
Objective 1 only describes levels, so there is no test and no hypothesis. Objectives 2, 3 and 4 each need a test, and each one gets its own hypothesis. The table at the end of this section gives them in the null form. Most Indian theses in management, commerce, education and social sciences write them this way.
Null hypothesis, H0, is the “nothing is there” line. In H0 you write that no difference or no relationship exists, and the alternative (H1) is saying something is there. For objective 2, on hostel and day students, you will write them like this.
- H0₁: There is no difference in sleep quality between hostel and day students.
- H1₁: There is a difference in sleep quality between hostel and day students.
Sometimes earlier studies give you a good reason to expect one direction. Then you can write a directional alternative, for example “Hostel students have poorer sleep quality than day students.” Check with your guide which form your department prefers. Many theses list only the null, and some list both.
Give a number to each hypothesis, so that you can refer to it in the results chapter, like “H0₂ is rejected (r = −.38, p < .001).” Keep the wording of each hypothesis exactly the same in Chapter 1, in the methodology chapter and in the results chapter. Examiners do compare them.
| Objective | Hypothesis (null form) | Likely test |
|---|---|---|
| 1. To assess the level of social media use and sleep quality among students | Not needed, as this objective only describes | Means, percentages, standard deviations |
| 2. To compare sleep quality between hostel and day students | H0₁: There is no difference in sleep quality between hostel and day students. | Independent-samples t-test, or Mann–Whitney U |
| 3. To examine the relationship between night-time social media use and sleep quality of these students | H0₂: There is no relationship between night-time social media use and sleep quality. | Pearson or Spearman correlation |
| 4. To identify which factors predict poor sleep quality | H0₃: Social media use, hostel stay, gender and course year do not predict sleep quality. | Multiple regression |
If your study is qualitative
Research using interviews, focus groups or case studies mostly does not test hypotheses. In such studies the scholar will frame research questions, and mostly these start with “how” or “why”. For example, “How do hostel students talk about using their phone at night?” or “Why do they stay online after midnight?”
Usually two to four research questions are enough. Keep them open, but not so open that any answer will fit. In a mixed methods study you may have objectives with hypotheses for the survey part and research questions for the interview part. Put a clear label on each one. Otherwise the committee member gets confused about which question the interviews are answering and which one the survey is answering.
Our research question sharpener gives quick feedback on a draft question, whether it is too broad, too narrow or too vague.
What committees ask about objectives
At the synopsis stage, most committee questions are on the objectives. Be ready for questions like these.
- “How will you measure this?” If you can’t say, the objective is still loose.
- “This hypothesis comes under which objective?” Each hypothesis should go back to one objective only.
- “Objective 3 and objective 4 look the same to me. What is the difference?”
- “You have written ‘impact’. Are you doing an experiment?”
In the viva also, many examiners keep results chapter open next to Chapter 1 and check the objectives one by one. Each objective needs an answer there, and each hypothesis a decision. In our research proposal and synopsis support we are working on the objectives and hypotheses along with the problem statement, so all of them match. At the analysis stage, the guide on choosing a statistical test matches each hypothesis to its test.
Questions scholars ask
Do I need both objectives and research questions?
Mostly no. Most Indian theses have objectives, with hypotheses for the testable ones. Some guides want research questions also. If you are giving both, every research question should match one objective.
Is “to give suggestions” a valid objective?
Most committees will not accept it. Suggestions come out of the findings, so their place is the final chapter. Still, in some departments the old habit is there and they want to see this line. What your department wants, your guide will know better than us. Talk to your guide before you drop it.
Should hypotheses be written in the null or alternative form?
In social sciences and management, Indian theses mostly list the null hypotheses. Science and medical theses many times state the expected effect directly. Look at recent theses from your department and follow what your guide says.
Can I add an objective after the synopsis?
Normally only with approval from your doctoral committee. Adding one quietly in the final thesis can create problems at evaluation, because the examiners compare the thesis with the approved synopsis.