Use a published standardised tool when a validated instrument for your variable already exists in the Indian population, adapt one when the construct fits but the language or age group does not, and construct your own only for achievement tests tied to a specific syllabus. Whichever route you take, your Chapter 3 must report the tool’s reliability and validity evidence and the permission you obtained to use it.
That is the whole decision, and most M.Ed scholars get it wrong in one direction: they build their own forty-item attitude scale from scratch because it feels like more original work, and then cannot produce validity evidence when the guide asks for it in the pre-submission review. Standardising a psychological instrument properly is a doctoral-scale project. Your dissertation is a one-year study. Use what exists.
What counts as a standardised tool in Indian education research?
A standardised tool is an instrument that has been administered to a defined norm group, scored by a fixed key, and published with a manual reporting its reliability coefficients, validity evidence and norms. The manual is what makes it standardised — not the fact that it appears in a published paper.
In Indian education departments, standardised tools typically come from three places:
- Commercial test publishers. Established Indian psychological test publishers, several of them based in Agra and Delhi, sell test booklets, manuals and scoring keys for attitude scales, adjustment inventories, study habits inventories, teaching aptitude tests, emotional intelligence scales and socio-economic status scales. You buy a set of booklets, usually priced per twenty-five or fifty copies, plus one manual.
- Tools published inside journal articles or theses. Many Indian scholars publish a scale as part of a paper, including the item pool and reliability figures. These can be used, but you must write to the author for permission and you must satisfy yourself that the reported psychometric work is real and not a single alpha coefficient with no validity evidence at all.
- Institutional and government instruments. Tools developed by NCERT, SCERTs, DIETs or the state examination boards, particularly for achievement testing and classroom observation.
When should you construct your own tool instead?
There is one clear case: an achievement test measuring learning of a specific unit of a specific syllabus. No published test can measure whether Class VIII students in your sample learnt the chapter you taught with your intervention, so you build it — and building it correctly is a recognised, examinable methodological contribution.
Constructing an achievement test in the Indian M.Ed tradition means producing a blueprint before you write a single item: a two-way table crossing content units against objectives — knowledge, understanding, application, skill — with the number of items and marks allotted in each cell. Then you write items to fill the cells, get the draft reviewed by subject experts and teachers for content validity, pilot it on a group similar to but excluded from your final sample, and run item analysis to compute the difficulty index and discrimination index for each item. Items that everyone answers correctly, or that fail to discriminate between high and low scorers, are dropped or revised. The final test, its blueprint, the item analysis table and the reliability estimate all go into the appendix.
The second, weaker case for building your own tool is a construct so specific to your context that nothing exists — a scale on teachers’ attitudes towards a state-specific policy, for instance. Discuss it with your guide before committing, because you are taking on the burden of demonstrating validity yourself.
How do you choose between three tools measuring the same construct?
Suppose three published Indian scales all claim to measure academic achievement motivation. Compare them on five things, in this order:
- Norm group. Was the tool standardised on the population you are studying — same age band, same school stage, same broad linguistic region? A scale normed on urban college students is a poor fit for rural secondary school pupils, and the examiner will say so.
- Reported reliability. The manual should give a test-retest coefficient, a split-half coefficient or Cronbach’s alpha, with the sample it was computed on. A manual that reports a coefficient without a sample size is not evidence.
- Validity evidence. Content validity through expert judgement is the minimum; criterion or construct validity against another measure is better.
- Language and length. A ninety-item English inventory administered to Class IX students in a regional-medium school will produce fatigue effects, not data.
- Availability and cost. Can you actually buy it, in the quantity you need, within your timeline? Booklets are consumable and priced per copy, so a sample of 300 costs real money.
Write this comparison into your dissertation. A short paragraph explaining why you selected one tool over two named alternatives is one of the cheapest ways to demonstrate methodological judgement to an examiner.
Do you need permission to use a published tool?
Yes, and this is where M.Ed scholars most often cut a corner that surfaces later. Published tests are copyrighted. Purchasing the booklets from the publisher is itself the licence for commercial instruments. For a tool published in a journal or a thesis, you write to the author, describe your study, and ask for written permission to use — and, if applicable, to translate — the instrument. Email is fine. Keep the reply and put it in your appendix.
Photocopying a purchased test booklet to avoid buying copies is a copyright violation, and some universities treat it as a research misconduct matter. It is also visible: photocopied booklets appear in your appendix.
How do you adapt or translate a tool into a regional language?
Adaptation is legitimate and common in Indian education research, because a national sample rarely shares one medium of instruction. It is also the step that generates the most examiner questions, so document it properly:
- Obtain the author’s or publisher’s permission to translate.
- Have the instrument translated into the target language by a bilingual expert in the subject, not by a general translator.
- Have a second independent bilingual expert back-translate into the original language without seeing the source.
- Compare the back-translation with the original, and resolve discrepancies item by item with a panel.
- Pilot the translated version on a small group from the target population — typically 25 to 50 respondents who will not appear in the final sample.
- Re-establish reliability on the translated version and report the coefficient for your own data.
That last step is not optional. A reliability coefficient from the original manual does not transfer to your translation or to your sample. Report the alpha you computed yourself, and for guidance on what value is defensible and how to diagnose a low one, see our explanation of what counts as an acceptable Cronbach’s alpha.
What must the tools section of Chapter 3 contain?
For each instrument, in this order: the full name of the tool with its author and year; what it measures and how it is structured, including number of items, dimensions and response format; the scoring procedure with any reverse-scored items; the reliability and validity evidence from the manual with the norm group named; the reliability you computed on your own sample; the permission obtained; and the administration procedure including time taken and conditions.
Here is a worked example of the level of specificity expected:
“Teacher attitude towards inclusive education was measured using a published five-point Likert-type attitude scale comprising 40 statements across four dimensions, 14 of which are negatively worded and reverse scored. The manual reports a split-half reliability of the order commonly regarded as adequate for group comparison, computed on a norm group of secondary school teachers. Written permission to use and to translate the scale into the regional medium was obtained from the author and is reproduced in Appendix C. The translated version was back-translated by an independent bilingual expert and piloted on 40 teachers not included in the final sample; Cronbach’s alpha computed on the present sample (N = 320) was 0.87. Administration took approximately 25 minutes per respondent, in school, during a free period arranged with the head teacher.”
Notice what makes that paragraph work: named dimensions, reverse-scored item count, the permission located in a numbered appendix, the alpha computed on the present sample, and a realistic administration note. Nothing is claimed that cannot be shown.
How many respondents does your tool actually need?
Sample size in an M.Ed dissertation is driven by the analysis, not by a departmental rule of thumb. A study comparing two groups on one scale needs far fewer respondents than one running a factor analysis on a newly constructed instrument. Work out the requirement before data collection using the approach set out in our guide to calculating sample size with G*Power and the Krejcie-Morgan table, and record the assumptions you used, because the viva question is almost always “why this number?”.
Remember that booklets are consumable. Deciding you need 400 respondents rather than 250 after you have bought the test sets is an expensive correction.
What clearances do you need before administering the tool?
Education research in India runs on institutional permission more than on formal ethics committees, but both may apply. You will normally need written permission from the school head or the block education officer, informed consent from adult participants, and parental consent where respondents are minors — which covers most school-based M.Ed studies. Where your university operates a research ethics committee, or where your study touches on sensitive topics such as mental health, disability or caste, formal clearance is required before collection begins, not afterwards; our walkthrough of obtaining ethics committee clearance for an Indian thesis covers the paperwork and the usual timelines.
How do you report tool-based results in Chapter 4?
Report the instrument’s descriptive statistics before any inferential test — mean, standard deviation and N for each group, dimension by dimension where the tool has dimensions. Give the reliability figure once, in Chapter 3, not again in every table. Where the tool has established norms, interpret group scores against those norms as well as against each other, because that is the advantage of using a standardised instrument in the first place. The conventions for laying out those tables are covered in our guide to writing the results chapter with APA tables and figures.
If your study also tests relationships between several measured constructs, the model linking them belongs in Chapter 2, not Chapter 4 — see how to build and justify a conceptual framework, which applies to education research as directly as it does to management.
Get Chapter 3 written while the data collection is still running
The tools section, the sampling description and the procedure paragraph can all be drafted before your last questionnaire comes back — they describe decisions you have already made. Tesify turns your notes on the instrument, the permissions and the pilot into a properly structured methodology chapter in your department’s format, with citations built as you go, so the writing is not waiting on the data.
Draft your methodology chapter with Tesify and stop losing weeks to the parts you already know.
Frequently asked questions
Can I use a standardised tool I found in someone else’s thesis on Shodhganga?
You can use it as a lead, but not as your source. Find the original publication of the tool, verify the psychometric evidence in the manual or original paper, and obtain permission from the tool’s author — not from the scholar whose thesis you found it in. Reproducing a scale from another thesis without tracing it to its author is both a copyright and an attribution problem.
Is a self-constructed questionnaire acceptable in an M.Ed dissertation?
For factual information — age, teaching experience, school type, resources available — yes, a self-constructed information schedule is standard and needs no standardisation. For measuring a psychological construct such as attitude, motivation or adjustment, a self-constructed instrument requires you to demonstrate content validity through expert judgement, run a pilot, perform item analysis and report reliability. Departments accept it when that work is shown.
What is the difference between a tool and a technique in education research?
The technique is the method of data collection — questionnaire, interview, observation, testing. The tool is the specific instrument used to carry out that technique, such as a named attitude scale or an observation schedule. Indian M.Ed dissertations conventionally have separate sub-sections for each, and confusing them is a common examiner comment.
How many items should an achievement test have?
The blueprint decides, not a fixed number. Allocate items proportionally to the weight of each content unit and each objective, then write roughly one and a half times as many items as you need so that item analysis can eliminate the weak ones. A unit test for a single chapter commonly ends up between 25 and 50 objective items after item analysis.
What are the difficulty index and discrimination index?
The difficulty index is the proportion of pilot respondents who answered an item correctly; items answered by almost everyone or almost nobody carry little information. The discrimination index compares performance on the item between the high-scoring and low-scoring groups; an item that the weaker group answers as often as the stronger group is not measuring the construct. Both are computed from pilot data and reported in a table in your appendix.
Do I have to pay for standardised test booklets myself?
Usually yes, though some departments hold copies of commonly used instruments in the education laboratory or library and permit scholars to purchase only the consumable answer sheets. Ask your guide and the department office before ordering — and ask early, because delivery of test materials can take weeks and cannot be compressed at the end of the year.
Should I report the manual’s reliability or my own?
Both. The manual’s coefficient establishes that the instrument is sound in its norm group; your own coefficient establishes that it behaved acceptably in your sample, which is what your findings actually rest on. Report the manual’s figure in the description of the tool and your own immediately afterwards, with your sample size.
Can I shorten a standardised scale to save administration time?
Dropping items breaks the standardisation — the norms and the reported reliability no longer apply, and the scores are no longer comparable to the manual’s norm tables. If length is genuinely unworkable, look for a published short form of the same instrument, which its author may have validated separately, rather than cutting items yourself.
Does the tool have to appear in the appendix of the dissertation?
Self-constructed tools must be reproduced in full. For copyrighted commercial instruments, many publishers permit only a sample of items to be shown, and your appendix should then carry the permission letter, the tool’s front page and a few illustrative items with a note explaining the restriction. Check what your university ordinance requires before submission.