iGROW PsyCapital
Technical manual. What this assessment measures, how well it measures it, and exactly where its limits are.
Photo: www.kaboompics.com / PexelsContents
What is in this manual
- 01What this manual is
- 02What PsyCapital measures
- 03Where it comes from
- 04The questions people answer
- 05How scoring works
- 06Does it measure consistently?
- 07Do the questions belong together?
- 08What typical scores look like
- 09The two risk questions
- 10Where this assessment applies
- 11Scope of the current evidence
- 12Using the results well
- 13Methods and data
- 14The science it is built on
- 15About iGROW
HR leaders, wellbeing sponsors and the psychologists who advise them. You do not need a statistics background to read it. Where a technical idea matters, it is explained in plain words first and the number follows.
What this manual is
PsyCapital is a short assessment of psychological wellbeing at work. It asks 19 questions, takes about seven minutes, and reports an organisation's results as group patterns rather than individual diagnoses.
This manual exists because buyers should be able to check our work. It sets out what the assessment measures, where the questions came from, how the scoring runs, what the evidence says about its quality, and which claims it cannot support. We have included the weaknesses alongside the strengths. An assessment that only publishes its good numbers is not telling you enough to make a decision.
The honest frame, before anything else
PsyCapital measures how people describe their own wellbeing at a point in time. It is a self-report questionnaire, so it captures perception. That is genuinely useful for spotting where support is needed and for tracking change across an organisation. It is not a clinical diagnosis, and it describes how people feel now rather than forecasting what they will do next. Section 10 sets out where it applies, and section 11 covers the evidence behind it.
The 15 wellbeing questions hold together with a consistency score of 0.89. In plain terms, the questions behave as though they are measuring the same underlying thing, which is what you want before you trust a total score. The usual bar for reporting group results is 0.70.
What PsyCapital measures
Seven factors describe what is going well, and two further questions check for signs of distress. Together they give an organisation a picture of both sides of wellbeing.
Whether someone trusts their own judgement and holds to what they think matters, even when others disagree.
Whether work connects to something the person finds meaningful.
The everyday balance of good feeling against bad, and how someone views their life so far.
Whether close, warm and trusting relationships are present or hard to sustain.
Whether someone is using their strengths and still growing rather than standing still.
Whether daily demands feel manageable and within the person's control.
Where it comes from
Whether colleagues have your back and good work gets noticed.
The wellbeing questions are adapted from established academic measures of psychological wellbeing, drawn from the research literature on what allows people to function well rather than merely avoid illness. iGROW adapted that work for workplace use, shortened it to a length employees will actually complete, and has refined the wording through successive deployments since 2009.
How it developed
The first version, called the iGROW HoneyComb Measure, ran from 2009 as an 18-item questionnaire. In 2014 and 2015 we ran a development study on a sample of Singapore working adults, mapping our own factors against several established wellbeing frameworks and testing which items performed best. The current form, PsyCapital, came into use in 2021. It kept the strongest wellbeing items, added questions about purpose and support at work, and replaced the earlier distress questions with two standard clinical screening questions used worldwide.
Questions that have been tested by researchers across many samples carry evidence with them. Writing fresh questions means starting that evidence from zero. Building on well-established academic work means the instrument inherits decades of prior research rather than asking you to take our word for it.
The two risk questions
The distress questions are the PHQ-2 and the GAD-2, short screening questions for low mood and for anxiety. They are in the public domain, are used in primary care around the world, and have been tested extensively. We use them word for word. Changing the wording would break the comparison with published research, which is the main reason to use them at all.
The questions people answer
Nineteen questions in total. Fifteen cover the seven wellbeing factors, two ask about low mood, and two ask about anxiety. Two optional open comment boxes follow, and then a short set of background questions.
The wellbeing questions
People move a slider from 0 to 10, where 0 means strongly disagree and 10 means strongly agree. Every factor is measured by two questions, except Purpose in Life, which uses three.
Strongly disagree on the left, strongly agree on the right. An answer must be given, so the middle position is not recorded unless the person chooses it.
Two questions are worded backwards on purpose
The two relationship questions describe difficulty rather than success, for example “I have not experienced many warm and trusting relationships with others”. Agreeing with those indicates lower wellbeing, so their scores are reversed before the factor is calculated. Mixing question direction like this is standard practice. It stops people slipping into a habit of ticking the same end of every scale without reading.
The risk questions
These ask how often something has been true over the past two weeks, with four options running from “not at all” to “nearly every day”. The wording is fixed by the original clinical instruments and we do not alter it.
PsyCapital exists in English, Thai, Bahasa Indonesia, Bahasa Malaysia and Mandarin. Translation quality matters more than it might appear, and section 11 describes a specific translation problem we found in the Thai version of one question during a 2026 review of our own data.
How scoring works
The arithmetic is deliberately simple in principle. Each factor is calculated from the average of its questions on the 0 to 10 scale, shown as a percentage when a report needs one. Nothing is weighted, and no factor counts for more than another.
The seven factors can be averaged into one number, and statistically that total behaves well. We do not lead with it, because a single figure hides the thing you need. Two organisations can sit at the same overall score while one has an engagement problem and the other has a relationships problem, and those need different responses.
The risk questions
Each of the two questions scores 0 to 3, so each pair totals between 0 and 6. A total of 3 or more is the recognised threshold at which further conversation is warranted. This threshold comes from the published clinical research, not from us. We report the percentage of people at or above it.
Someone scoring at or above the threshold is not being diagnosed with anything. The screening questions are a prompt to look further, nothing more. In every deployment, anyone whose answers suggest distress is shown support information immediately, regardless of what the organisation sees in its aggregate report. That routing is built and monitored by our team as a standing part of the service. An organisation collecting these same two questions on its own takes on that duty of care itself, including the responsibility to notice and act on a flagged answer.
Does it measure consistently?
Consistency asks a simple question. If several questions are meant to measure the same thing, do people answer them in a related way? A score of 0 means no relationship at all and 1 means perfect agreement. For reporting group results, 0.70 and above is the usual expectation.
Each factor on its own
| Factor | Score | |
|---|---|---|
| Purpose in Life | .89 | |
| Environmental Mastery | .80 | |
| Positive Relationships | .77 | |
| Engagement | .74 | |
| Support at Work | .74 | |
| Positive Emotions | .73 | |
| Realistic Optimism | .67 | weakest |
Realistic Optimism sits at 0.67, below the 0.70 expectation. Its two questions are related but not as tightly as the rest, which means a score on this factor alone should be read with more caution than the others. This was flagged in our own development study back in 2015 and it remains the case. The fix is to add one or two further questions to that factor in the next revision, which is planned rather than hypothetical.
Do the questions belong together?
The second check asks whether the groupings are real. We put two questions together and call them Engagement, but do people's answers actually cluster that way, or have we imposed a structure that is not there?
To test this we set the labels aside and let a statistical procedure look for natural clusters in the answers of all 1,854 people. The technique is called factor analysis. It does not know what we named anything. It simply looks at which questions move together.
What it found
Two things, and both are worth stating. First, there is one broad underlying pattern running through all fifteen questions. People who feel good in one area tend to feel good in others, which is a normal and well-documented finding in wellbeing research. Second, when we asked the procedure to sort the questions into seven groups, it reproduced our seven factors almost exactly. Each pair of questions we had grouped together did land together.
The seven factors are best understood as different views of a connected picture rather than seven independent traits. That is why a report showing one factor far below the others is informative. It is telling you something specific has broken away from a person's or a team's general pattern, and that is usually where attention is worth spending.
A correction we made to our own scoring
During the 2026 review we found that two of the seven factors had been described incorrectly in an internal planning document, which assigned one question to the wrong factor. The assessment itself was always scored correctly. We checked the live questionnaire and reproduced the original published figures to confirm it. We mention it because the correction is part of the record, and because the check is the kind of thing a manual should be able to show it has done.
What typical scores look like
These are the average scores across all 1,859 employees in the evidence base. They give a client organisation something to compare against, which is more useful than a score sitting on its own.
| Factor | Average out of 10 | As a percentage |
|---|---|---|
| Engagement | 8.1 | 81% |
| Environmental Mastery | 7.6 | 76% |
| Positive Emotions | 7.2 | 72% |
| Support at Work | 7.1 | 71% |
| Purpose in Life | 6.9 | 69% |
| Realistic Optimism | 6.8 | 68% |
| Positive Relationships | 6.0 | 60% |
Based on 1,854 people who answered every wellbeing question. Detailed reference tables, including breakdowns by country, gender, age and length of service, are supplied with client reports rather than published here.
Reading the pattern
Engagement scores highest and Positive Relationships lowest, and that ordering has held across every group we have measured since 2021. The gap is wide. People report using their strengths at work far more readily than they report having warm and trusting relationships. For most organisations we work with, the relationship result is the more useful place to start.
Younger employees in this sample report the lowest wellbeing and the highest distress, and both improve steadily with age. The youngest group scores around 68% overall against 73% for those over 50. If your organisation skews young, expect that to show, and compare within age bands rather than against the overall average alone.
The two risk questions
Alongside wellbeing, PsyCapital carries two short screening questions for low mood and two for anxiety. These reference figures need more care than the wellbeing ones, and we would rather explain why than print a single tidy number.
of employees scored at or above the anxiety threshold. This figure holds steady across every language we measured, so it travels well.
of employees scored at or above the low-mood threshold. This one is not steady across languages and should not be quoted on its own.
Why the low-mood figure carries a warning
Seven in ten people in this evidence base answered in Thai, and during the 2026 review we found that one of the two Thai low-mood questions does not behave like its English or Indonesian counterparts. People endorse it far more readily, and it does not track with the second question the way it should. That points to the translation rather than to Thai employees, particularly because the two anxiety questions in the same Thai questionnaire behave perfectly normally.
The practical effect is that the 30.6% figure is pulled upward by a translation artefact. Without the Thai responses the same calculation gives 21%. We are reporting both rather than quietly choosing one.
Low-mood results are reported by country rather than as one regional figure, Thai low-mood results are labelled as indicative, and the Thai wording is being reviewed against the published Thai version of the original clinical questionnaire. Anxiety results are unaffected and need no such caveat. We would rather hand you a figure with a caution attached than a cleaner number you cannot rely on.
Where this assessment applies
Every assessment is built for a particular job. Setting out what PsyCapital is designed to do, and what it should not be asked to do, is what lets you rely on it inside that scope.
It is not a clinical diagnosis
PsyCapital does not diagnose depression, anxiety or any other condition, and no score it produces should ever be recorded as though it did. The two screening questions indicate that a conversation may be worthwhile. Diagnosis requires a qualified clinician and a proper assessment.
It is not a performance or selection tool
Results must never be used to hire, promote, manage out, or appraise anyone. The assessment was not designed for those decisions and there is no evidence it would be fair or accurate in them. Using a wellbeing measure this way also destroys the honesty of the answers, and with it the value of the exercise.
It describes the present rather than forecasting the future
PsyCapital gives you an accurate reading of how a workforce is doing now, which is what it was designed for. Linking those readings to later outcomes such as turnover or absence requires following the same people over time, and that study is our current development priority. Until it is complete, treat the results as a clear picture of today rather than a forecast.
It is not a before-and-after measure of your programme
Comparing your own organisation across two measurement rounds is sound practice and we encourage it. Attributing a change specifically to a training programme is a separate claim, and one that needs a study design built for it. We will tell you which of the two your data supports.
We publish these boundaries because a measure you can trust is one whose edges are visible. Inside this scope, PsyCapital does its job well, and that is a stronger claim than one made without limits.
Scope of the current evidence
Every assessment rests on a defined body of evidence. This page sets out what has been established for PsyCapital, what the reference figures are drawn from, and what we are building next.
What is established
What the reference figures are drawn from
The current figures come from deployments between 2021 and 2023 across Thailand, Indonesia and Singapore, with Thai-language respondents forming the largest group. Because the mix of countries shapes any regional average, we report results by country as well as overall, and recommend comparing against the country cut closest to your own workforce.
What we are building next
- Predictive evidenceFollowing the same employees over time and comparing scores against outcomes such as absence and retention. This is our main development priority.
- Score stabilityMeasuring the same people twice a short interval apart, to establish how much an individual score moves on its own.
- Realistic OptimismAdditional questions for the one factor sitting below the 0.70 expectation, scheduled for the next revision.
- Thai wordingReview of one low-mood question, described in section 9, against the published Thai source questionnaire.
PsyCapital asks people to describe their own experience, which is true of every workplace survey and is why we report group patterns rather than individual conclusions. Within that frame the measurement is sound, and the patterns have held across every organisation we have measured since 2021.
Using the results well
A few practical rules, some of which reduce what we can sell you. They are here because they are what makes the exercise work.
Protect anonymity, and be seen to
No group smaller than five people is ever reported separately, no matter how the results are sliced. If employees suspect their answers can be traced back, they will answer carefully rather than honestly, and the results stop being worth collecting. Say this clearly when you invite people to take part.
Read the shape, not the single number
The useful signal is the pattern. Which factor sits furthest below the others? Does it differ by department, by age band, by length of service? An organisation at 71% overall with a relationships score of 55% has a specific and addressable problem. The overall figure alone would not have told you that.
Decide what you will do before you ask
The fastest way to damage trust is to run a wellbeing survey and then do nothing visible with it. Agree in advance what you will share with employees and roughly when. People are markedly more willing to answer honestly the second time if something happened after the first.
In most organisations we measure, the lowest scoring factor is Positive Relationships, and the strongest response to it is not a training programme. It is ordinary management practice. Regular one to one time that is not about task status, teams that eat together, and workloads that leave room for a conversation. If your results point there, start with how work is organised. Bring in outside help only if that is not enough.
Measure again, and leave enough time
Wellbeing moves slowly. Twelve months between measurements is sensible for most organisations, and six is the shortest interval we would suggest. Anything more frequent tends to measure the mood of the week.
What stays with us between waves
The questions are short and deliberately simple. Harder to reproduce: a benchmark from years of real deployments, a monitored path for flagged answers, and translations checked over time. Measuring alone starts from zero on all three, and means owning that duty of care too.
Methods and data
Evidence base
1,859 employees completed PsyCapital across five deployments between December 2021 and October 2023. Respondents answered in Thai (1,306), English (351), Bahasa Indonesia (180), and smaller numbers in Bahasa Malaysia and Mandarin. Statistical results use the 1,854 people who answered every wellbeing question.
Instrument
Fifteen wellbeing questions answered on a 0 to 10 slider, covering seven factors. Two questions are reverse worded. Four screening questions covering low mood and anxiety, answered on a four point frequency scale, used with their original published wording. Background questions cover gender, age band, length of service and role.
Analysis
Consistency was assessed for each factor and for the full set of wellbeing questions. Structure was examined by factor analysis across all fifteen questions, with the number of underlying patterns determined by comparison against randomly generated data rather than chosen by us. Screening results use the published clinical threshold of 3 or more. Reference figures are reported overall and separately by country, gender, age band and length of service. All calculations were rerun from the original response files in 2026 rather than carried forward from earlier summaries, and the published 2023 figures were reproduced exactly as a check.
Ethics and data protection
Participation is voluntary and results are reported only in aggregate, with a minimum group size of five. Personal identifiers are removed before any analysis. Data is handled under prevailing personal data protection requirements in the countries where the assessment runs. The wellbeing questions are adapted from established academic measures of psychological wellbeing, and the screening questions are in the public domain and used in their original published wording. The research this instrument draws on is listed on the following page.
Corrections in this edition
Two figures published in an earlier internal summary have been withdrawn. They described distress levels in a way that could not be reproduced from the underlying responses and that mixed two different measurement conventions. The figures in this manual replace them, and the calculation behind each is documented.
The science it is built on
PsyCapital draws on established, published research in psychology. The wellbeing questions build on the science of what allows people to function well, and the two screening questions are public-domain clinical tools used worldwide. The main sources are listed here so any reviewer can go to the original work.
- Ryff, C. D. (1989). Happiness is everything, or is it? Explorations on the meaning of psychological well-being. Journal of Personality and Social Psychology, 57(6), 1069–1081. Wellbeing model
- Ryff, C. D., & Keyes, C. L. M. (1995). The structure of psychological well-being revisited. Journal of Personality and Social Psychology, 69(4), 719–727. Wellbeing model
- Seligman, M. E. P. (2011). Flourish: A visionary new understanding of happiness and well-being. Free Press. Positive psychology
- Steger, M. F., Dik, B. J., & Duffy, R. D. (2012). Measuring meaningful work: The Work and Meaning Inventory (WAMI). Journal of Career Assessment, 20(3), 322–337. Purpose at work
- Schaufeli, W. B., Bakker, A. B., & Salanova, M. (2006). The measurement of work engagement with a short questionnaire. Educational and Psychological Measurement, 66(4), 701–716. Engagement
- Kroenke, K., Spitzer, R. L., & Williams, J. B. W. (2003). The Patient Health Questionnaire-2: Validity of a two-item depression screener. Medical Care, 41(11), 1284–1292. Low-mood screener
- Kroenke, K., Spitzer, R. L., Williams, J. B. W., Monahan, P. O., & Löwe, B. (2007). Anxiety disorders in primary care: Prevalence, impairment, comorbidity, and detection. Annals of Internal Medicine, 146(5), 317–325. Anxiety screener
Citing this research is an acknowledgement of the science the instrument is grounded in. The specific items, scoring and benchmarks in PsyCapital are the work of iGROW Pte Ltd.
iGROW Pte Ltd
iGROW is an award winning psychological consultancy based in Singapore, working across Southeast Asia since 2009. Our work covers workplace mental health, wellbeing measurement, and the training and coaching that follows from what the measurement shows.
Photo: Mikhail Nilov / PexelsThese are firm wide figures covering all our work since 2009. They are separate from the 1,859 employees in this manual's evidence base, and the two should not be combined.
Questions about this manual
We are glad to walk any client, adviser or reviewer through the calculations behind these figures, including the parts we have flagged as weak. Write to psychologists@igrow.co.
iGROW Pte Ltd · Singapore · igrow.co
iGROW PsyCapital and the iGROW HoneyComb Measure are proprietary instruments of iGROW Pte Ltd, adapted from established academic measures of psychological wellbeing. Technical manual, edition 1.0, 2026. This manual may be shared in full. Please quote figures from it together with the accompanying notes on scope. No part may be resold or repackaged without written consent.